diff --git a/experiments/h20/dsv4_h20_sglang_tp_dp_matrix/run_adaptive_concurrency.sh b/experiments/h20/dsv4_h20_sglang_tp_dp_matrix/run_adaptive_concurrency.sh index 8f48c5f..7972a9e 100755 --- a/experiments/h20/dsv4_h20_sglang_tp_dp_matrix/run_adaptive_concurrency.sh +++ b/experiments/h20/dsv4_h20_sglang_tp_dp_matrix/run_adaptive_concurrency.sh @@ -6,15 +6,15 @@ SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" EXPERIMENT_NAME="$(basename "$SCRIPT_DIR")" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/lib.sh" +source "${SCRIPT_DIR}/../../../scripts/common/lib.sh" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/platform.sh" +source "${SCRIPT_DIR}/../../../scripts/common/platform.sh" # shellcheck source=/dev/null source "${SCRIPT_DIR}/config.env" # shellcheck source=/dev/null source "${SCRIPT_DIR}/adaptive_config.env" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/adaptive_bench_lib.sh" +source "${SCRIPT_DIR}/../../../scripts/common/adaptive_bench_lib.sh" ENGINE="sglang" ENGINE_PORT="$SGLANG_PORT" diff --git a/experiments/h20/dsv4_h20_sglang_tp_dp_matrix/run_adaptive_concurrency_add16.sh b/experiments/h20/dsv4_h20_sglang_tp_dp_matrix/run_adaptive_concurrency_add16.sh index d7b26a0..c98b531 100755 --- a/experiments/h20/dsv4_h20_sglang_tp_dp_matrix/run_adaptive_concurrency_add16.sh +++ b/experiments/h20/dsv4_h20_sglang_tp_dp_matrix/run_adaptive_concurrency_add16.sh @@ -6,15 +6,15 @@ SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" EXPERIMENT_NAME="$(basename "$SCRIPT_DIR")" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/lib.sh" +source "${SCRIPT_DIR}/../../../scripts/common/lib.sh" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/platform.sh" +source "${SCRIPT_DIR}/../../../scripts/common/platform.sh" # shellcheck source=/dev/null source "${SCRIPT_DIR}/config.env" # shellcheck source=/dev/null source "${SCRIPT_DIR}/adaptive_config.env" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/adaptive_bench_lib.sh" +source "${SCRIPT_DIR}/../../../scripts/common/adaptive_bench_lib.sh" ENGINE="sglang" ENGINE_PORT="$SGLANG_PORT" diff --git a/experiments/h20/dsv4_h20_sglang_tp_dp_matrix/run_bench.sh b/experiments/h20/dsv4_h20_sglang_tp_dp_matrix/run_bench.sh index 948aa52..3fbb7b3 100755 --- a/experiments/h20/dsv4_h20_sglang_tp_dp_matrix/run_bench.sh +++ b/experiments/h20/dsv4_h20_sglang_tp_dp_matrix/run_bench.sh @@ -6,9 +6,9 @@ SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" EXPERIMENT_NAME="$(basename "$SCRIPT_DIR")" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/lib.sh" +source "${SCRIPT_DIR}/../../../scripts/common/lib.sh" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/platform.sh" +source "${SCRIPT_DIR}/../../../scripts/common/platform.sh" # shellcheck source=/dev/null source "${SCRIPT_DIR}/config.env" @@ -524,7 +524,7 @@ run_parallel_config() { # Parse results. log "parsing ${config_label} results" - "$PYTHON" "${SCRIPT_DIR}/../../../../scripts/common/parse_backend.py" "$result_root" --backend sglang \ + "$PYTHON" "${SCRIPT_DIR}/../../../scripts/common/parse_backend.py" "$result_root" --backend sglang \ >> "${phase_log_dir}/parse.log" 2>&1 || { log "WARNING: parser failed for ${config_label}; see ${phase_log_dir}/parse.log" } diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c112_i1024_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c112_i1024_o2048_a1.json new file mode 100644 index 0000000..cbfe5b3 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c112_i1024_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 560, + "failed": 0, + "duration_s": 271.93914603099984, + "request_tps": 2.059284248602305, + "input_tps": 2108.70707056876, + "output_tps": 4217.41414113752, + "total_tps": 6326.121211706281, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 2360.8657174918335, + "ttft_p95_ms": 4785.738604894141, + "ttft_p99_ms": 5083.200499021332, + "tpot_p50_ms": 25.078506466781775, + "tpot_p95_ms": 26.261017000906556, + "tpot_p99_ms": 26.82069243732855, + "e2e_p50_ms": 54802.16980399564, + "e2e_p95_ms": 55957.59202675035, + "e2e_p99_ms": 55983.62555622443, + "itl_p50_ms": 24.729957003728487, + "itl_p95_ms": 26.741590403253216, + "itl_p99_ms": 29.055896277423017, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i1024_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i1024_o1024_a1.json new file mode 100644 index 0000000..a45c7e1 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i1024_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 82.26222992698604, + "request_tps": 0.9724997738452514, + "input_tps": 995.8397684175375, + "output_tps": 995.8397684175375, + "total_tps": 1991.679536835075, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 544.7934150070068, + "ttft_p95_ms": 881.7404834531771, + "ttft_p99_ms": 897.7118983241962, + "tpot_p50_ms": 15.400015403226577, + "tpot_p95_ms": 15.83289659242619, + "tpot_p99_ms": 16.04854842089475, + "e2e_p50_ms": 16476.469576002273, + "e2e_p95_ms": 16701.17062768695, + "e2e_p99_ms": 16705.017296750884, + "itl_p50_ms": 15.259214495017659, + "itl_p95_ms": 16.13290750174201, + "itl_p99_ms": 17.04316050017951, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i1024_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i1024_o128_a1.json new file mode 100644 index 0000000..69b9be5 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i1024_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 25.27050317999965, + "request_tps": 3.165746223182292, + "input_tps": 3241.724132538667, + "output_tps": 405.2155165673334, + "total_tps": 3646.9396491060006, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 1051.9641595019493, + "ttft_p95_ms": 7164.641508652858, + "ttft_p99_ms": 7186.009613565839, + "tpot_p50_ms": 18.72120111024348, + "tpot_p95_ms": 65.18447974448227, + "tpot_p99_ms": 66.09945481066043, + "e2e_p50_ms": 3493.628591997549, + "e2e_p95_ms": 9117.589514059364, + "e2e_p99_ms": 9138.18436251444, + "itl_p50_ms": 15.28507299371995, + "itl_p95_ms": 16.032998800073983, + "itl_p99_ms": 21.76707776481644, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i1024_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i1024_o2048_a1.json new file mode 100644 index 0000000..92046a3 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i1024_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 160.96535958899767, + "request_tps": 0.4970013436696486, + "input_tps": 508.92937591772017, + "output_tps": 1017.8587518354403, + "total_tps": 1526.7881277531606, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 408.2802780030761, + "ttft_p95_ms": 804.0736109498539, + "ttft_p99_ms": 871.2221907320782, + "tpot_p50_ms": 15.45796976209212, + "tpot_p95_ms": 15.706438855319314, + "tpot_p99_ms": 15.828019318818155, + "e2e_p50_ms": 32192.607425007736, + "e2e_p95_ms": 32558.569978662126, + "e2e_p99_ms": 32571.827183209243, + "itl_p50_ms": 15.350678993854672, + "itl_p95_ms": 16.041272398433648, + "itl_p99_ms": 16.693045923602767, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i1024_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i1024_o256_a1.json new file mode 100644 index 0000000..89da427 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i1024_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 25.495733696006937, + "request_tps": 3.137779871482159, + "input_tps": 3213.086588397731, + "output_tps": 803.2716470994327, + "total_tps": 4016.3582354971636, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 839.0470390004339, + "ttft_p95_ms": 1269.5357547512679, + "ttft_p99_ms": 1545.07869652909, + "tpot_p50_ms": 16.542784652961465, + "tpot_p95_ms": 18.45956203372243, + "tpot_p99_ms": 19.536147781054822, + "e2e_p50_ms": 5053.470186998311, + "e2e_p95_ms": 5453.463989494776, + "e2e_p99_ms": 5468.152181307523, + "itl_p50_ms": 15.194430503470358, + "itl_p95_ms": 15.815716297220208, + "itl_p99_ms": 16.660370378667714, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i1024_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i1024_o4096_a1.json new file mode 100644 index 0000000..008271a --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i1024_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 316.64033691798977, + "request_tps": 0.2526525861445129, + "input_tps": 258.7162482119812, + "output_tps": 1034.864992847925, + "total_tps": 1293.581241059906, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 575.7432910031639, + "ttft_p95_ms": 887.6251749068615, + "ttft_p99_ms": 907.8684282244649, + "tpot_p50_ms": 15.333096270205825, + "tpot_p95_ms": 15.443372411818606, + "tpot_p99_ms": 15.483845902724966, + "e2e_p50_ms": 63438.04139450367, + "e2e_p95_ms": 63812.25281149527, + "e2e_p99_ms": 63831.44774455199, + "itl_p50_ms": 15.270748001057655, + "itl_p95_ms": 15.853451890870927, + "itl_p99_ms": 16.49769142619334, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i1024_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i1024_o512_a1.json new file mode 100644 index 0000000..e37d0f6 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i1024_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 63.555366341999616, + "request_tps": 1.2587450061967969, + "input_tps": 1288.95488634552, + "output_tps": 644.47744317276, + "total_tps": 1933.4323295182799, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 651.5386349929031, + "ttft_p95_ms": 16920.61746209365, + "ttft_p99_ms": 16942.941599875485, + "tpot_p50_ms": 15.848476987276518, + "tpot_p95_ms": 23.578188225724393, + "tpot_p99_ms": 29.306536829046802, + "e2e_p50_ms": 8805.075798496546, + "e2e_p95_ms": 24636.574015000224, + "e2e_p99_ms": 24650.812979813927, + "itl_p50_ms": 15.365997998742387, + "itl_p95_ms": 16.487579100066792, + "itl_p99_ms": 18.37696666130795, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i131072_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i131072_o128_a1.json new file mode 100644 index 0000000..4c516eb --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i131072_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 653.2824726870022, + "request_tps": 0.12245851273332609, + "input_tps": 16050.882180982517, + "output_tps": 15.674689629865739, + "total_tps": 16066.556870612381, + "mean_input_tokens": 131072.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 55664.04897299799, + "ttft_p95_ms": 118594.41714475105, + "ttft_p99_ms": 139872.23895933607, + "tpot_p50_ms": 544.0846560747494, + "tpot_p95_ms": 902.7456552543442, + "tpot_p99_ms": 958.5754168834561, + "e2e_p50_ms": 129924.55754899856, + "e2e_p95_ms": 185606.83365028963, + "e2e_p99_ms": 243322.26077544419, + "itl_p50_ms": 16.123155495733954, + "itl_p95_ms": 1741.9904875576322, + "itl_p99_ms": 1853.2892793747305, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i131072_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i131072_o256_a1.json new file mode 100644 index 0000000..1130b45 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i131072_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 702.479465484008, + "request_tps": 0.11388233241078448, + "input_tps": 14926.785073746343, + "output_tps": 29.153877097160827, + "total_tps": 14955.938950843505, + "mean_input_tokens": 131072.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 57824.69000250421, + "ttft_p95_ms": 127985.28046924763, + "ttft_p99_ms": 151725.3745295175, + "tpot_p50_ms": 253.09320623725628, + "tpot_p95_ms": 525.102048641203, + "tpot_p99_ms": 619.6555398026641, + "e2e_p50_ms": 138529.73486200062, + "e2e_p95_ms": 184438.2270416012, + "e2e_p99_ms": 185996.91346610835, + "itl_p50_ms": 15.797007494256832, + "itl_p95_ms": 1590.6224329526603, + "itl_p99_ms": 1839.2413294677688, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i131072_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i131072_o512_a1.json new file mode 100644 index 0000000..2740038 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i131072_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 667.9211692050012, + "request_tps": 0.1197746136646944, + "input_tps": 15699.098162258824, + "output_tps": 61.32460219632353, + "total_tps": 15760.422764455148, + "mean_input_tokens": 131072.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 55981.32588399312, + "ttft_p95_ms": 121882.91273160065, + "ttft_p99_ms": 140760.88047753967, + "tpot_p50_ms": 141.8939732465815, + "tpot_p95_ms": 235.749314115342, + "tpot_p99_ms": 254.5075704539367, + "e2e_p50_ms": 132681.28659549984, + "e2e_p95_ms": 175247.86795944526, + "e2e_p99_ms": 234920.12947427476, + "itl_p50_ms": 15.690204003476538, + "itl_p95_ms": 1249.2242270018323, + "itl_p99_ms": 1797.7115144065467, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i16384_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i16384_o1024_a1.json new file mode 100644 index 0000000..de94686 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i16384_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 111.50969841099868, + "request_tps": 0.7174263865833329, + "input_tps": 11754.313917781326, + "output_tps": 734.6446198613329, + "total_tps": 12488.958537642658, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 4311.803800497728, + "ttft_p95_ms": 7437.313835108215, + "ttft_p99_ms": 9130.135508555395, + "tpot_p50_ms": 17.93213432404895, + "tpot_p95_ms": 22.04782108866086, + "tpot_p99_ms": 22.43329354472889, + "e2e_p50_ms": 21698.23511350114, + "e2e_p95_ms": 24644.28777230205, + "e2e_p99_ms": 26412.721827901405, + "itl_p50_ms": 15.224864997435361, + "itl_p95_ms": 15.731755192973651, + "itl_p99_ms": 16.860831878148, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i16384_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i16384_o128_a1.json new file mode 100644 index 0000000..026dbc0 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i16384_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 56.001648912992096, + "request_tps": 1.4285293657030234, + "input_tps": 23405.025127678335, + "output_tps": 182.851758809987, + "total_tps": 23587.87688648832, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 5949.235703003069, + "ttft_p95_ms": 9227.944630107957, + "ttft_p99_ms": 11097.429403630083, + "tpot_p50_ms": 39.72050276768026, + "tpot_p95_ms": 72.95558840476615, + "tpot_p99_ms": 80.88457267013912, + "e2e_p50_ms": 11065.538848997676, + "e2e_p95_ms": 13142.54083404594, + "e2e_p99_ms": 14229.989901815272, + "itl_p50_ms": 15.273823504685424, + "itl_p95_ms": 17.06440234847832, + "itl_p99_ms": 1069.5194568477748, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i16384_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i16384_o2048_a1.json new file mode 100644 index 0000000..b709aff --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i16384_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 188.87208026400185, + "request_tps": 0.4235671036617879, + "input_tps": 6939.723426394733, + "output_tps": 867.4654282993416, + "total_tps": 7807.1888546940745, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 2707.821463998698, + "ttft_p95_ms": 6159.3108133580245, + "ttft_p99_ms": 7476.327504011569, + "tpot_p50_ms": 16.789645155595252, + "tpot_p95_ms": 19.235649654987466, + "tpot_p99_ms": 19.809279338319755, + "e2e_p50_ms": 37331.52540450101, + "e2e_p95_ms": 41633.905589940696, + "e2e_p99_ms": 41644.779238444025, + "itl_p50_ms": 15.280855994205922, + "itl_p95_ms": 18.66166050749598, + "itl_p99_ms": 19.038321707921565, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i16384_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i16384_o256_a1.json new file mode 100644 index 0000000..2d5a40e --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i16384_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 64.32757946698985, + "request_tps": 1.2436345446676191, + "input_tps": 20375.708379834272, + "output_tps": 318.3704434349105, + "total_tps": 20694.07882326918, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 4847.543320000113, + "ttft_p95_ms": 8945.170914454502, + "ttft_p99_ms": 10200.10803171811, + "tpot_p50_ms": 30.697892317645408, + "tpot_p95_ms": 45.12670790233909, + "tpot_p99_ms": 46.21028031682417, + "e2e_p50_ms": 12845.963567502622, + "e2e_p95_ms": 15142.341455946007, + "e2e_p99_ms": 15201.770440481196, + "itl_p50_ms": 15.241901011904702, + "itl_p95_ms": 16.105174794211056, + "itl_p99_ms": 806.8968500866322, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i16384_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i16384_o512_a1.json new file mode 100644 index 0000000..5b0ad35 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i16384_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 72.41434827599733, + "request_tps": 1.1047534349835064, + "input_tps": 18100.28027876977, + "output_tps": 565.6337587115553, + "total_tps": 18665.914037481325, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 3977.608113505994, + "ttft_p95_ms": 6827.726796599017, + "ttft_p99_ms": 7064.793197970429, + "tpot_p50_ms": 20.6349563414749, + "tpot_p95_ms": 26.606408149222858, + "tpot_p99_ms": 27.689232515706706, + "e2e_p50_ms": 14308.850026493019, + "e2e_p95_ms": 14974.010983452172, + "e2e_p99_ms": 14998.68528622057, + "itl_p50_ms": 15.342150996730197, + "itl_p95_ms": 16.593942245526705, + "itl_p99_ms": 19.848780611646326, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i262144_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i262144_o128_a1.json new file mode 100644 index 0000000..e07d4d4 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i262144_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 1511.6757434350002, + "request_tps": 0.05292140219053523, + "input_tps": 13873.028055835668, + "output_tps": 6.77393948038851, + "total_tps": 13879.801995316056, + "mean_input_tokens": 262144.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 70474.41060249548, + "ttft_p95_ms": 206680.64425829216, + "ttft_p99_ms": 263259.9471526309, + "tpot_p50_ms": 1770.9045697204808, + "tpot_p95_ms": 1859.1631874913205, + "tpot_p99_ms": 1883.4251564144938, + "e2e_p50_ms": 293508.3537885075, + "e2e_p95_ms": 418159.4987228506, + "e2e_p99_ms": 491266.18717678235, + "itl_p50_ms": 1741.0211320020608, + "itl_p95_ms": 2297.1252548493794, + "itl_p99_ms": 2395.285054906126, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i262144_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i262144_o256_a1.json new file mode 100644 index 0000000..e595b4f --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i262144_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 1557.1239152909984, + "request_tps": 0.0513767717613209, + "input_tps": 13468.112456599705, + "output_tps": 13.15245357089815, + "total_tps": 13481.264910170603, + "mean_input_tokens": 262144.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 128086.70317550423, + "ttft_p95_ms": 293147.14115650055, + "ttft_p99_ms": 327246.22319525003, + "tpot_p50_ms": 694.1213535215974, + "tpot_p95_ms": 1110.7064755825186, + "tpot_p99_ms": 1121.9122485736023, + "e2e_p50_ms": 320114.9868900029, + "e2e_p95_ms": 442507.49014469614, + "e2e_p99_ms": 540593.0392589162, + "itl_p50_ms": 16.51875649986323, + "itl_p95_ms": 2135.694057152432, + "itl_p99_ms": 2348.4220405294036, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i4096_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i4096_o1024_a1.json new file mode 100644 index 0000000..c2df328 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i4096_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 84.46414080199611, + "request_tps": 0.9471475023648069, + "input_tps": 3879.516169686249, + "output_tps": 969.8790424215623, + "total_tps": 4849.395212107812, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 755.6758849968901, + "ttft_p95_ms": 1184.9343558955297, + "ttft_p99_ms": 1330.5567739978123, + "tpot_p50_ms": 15.835003849462446, + "tpot_p95_ms": 16.285198607422213, + "tpot_p99_ms": 16.55195564271857, + "e2e_p50_ms": 16962.706995996996, + "e2e_p95_ms": 17168.894911993266, + "e2e_p99_ms": 17178.149301250814, + "itl_p50_ms": 15.242625500832219, + "itl_p95_ms": 16.113438254251378, + "itl_p99_ms": 16.882939507922856, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i4096_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i4096_o128_a1.json new file mode 100644 index 0000000..d764008 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i4096_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 44.308171736003715, + "request_tps": 1.8055360188782963, + "input_tps": 7395.475533325502, + "output_tps": 231.10861041642193, + "total_tps": 7626.584143741924, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 1465.1815819961485, + "ttft_p95_ms": 23821.693100642005, + "ttft_p99_ms": 24259.730825575534, + "tpot_p50_ms": 22.75063312988532, + "tpot_p95_ms": 194.92663006346027, + "tpot_p99_ms": 197.04311093216822, + "e2e_p50_ms": 4635.849632992176, + "e2e_p95_ms": 26182.2721034383, + "e2e_p99_ms": 26191.62205229819, + "itl_p50_ms": 15.226024501316715, + "itl_p95_ms": 16.216387940949062, + "itl_p99_ms": 414.4044824814774, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i4096_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i4096_o2048_a1.json new file mode 100644 index 0000000..b696ba5 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i4096_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 161.3202662160038, + "request_tps": 0.4959079344245688, + "input_tps": 2031.2388994030339, + "output_tps": 1015.6194497015169, + "total_tps": 3046.858349104551, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 766.6918885006453, + "ttft_p95_ms": 1151.3519862077374, + "ttft_p99_ms": 1174.3669241876341, + "tpot_p50_ms": 15.32661797704204, + "tpot_p95_ms": 15.672662826941613, + "tpot_p99_ms": 15.77215861052261, + "e2e_p50_ms": 32366.54036600521, + "e2e_p95_ms": 32729.988757940737, + "e2e_p99_ms": 32742.016700837994, + "itl_p50_ms": 15.24984300340293, + "itl_p95_ms": 15.832478402444394, + "itl_p99_ms": 16.448479092650818, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i4096_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i4096_o256_a1.json new file mode 100644 index 0000000..428c75c --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i4096_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 32.527519440001925, + "request_tps": 2.4594559123256423, + "input_tps": 10073.931416885831, + "output_tps": 629.6207135553644, + "total_tps": 10703.552130441196, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 1533.099703505286, + "ttft_p95_ms": 2719.3758142995644, + "ttft_p99_ms": 2959.5566013005737, + "tpot_p50_ms": 19.392000529407206, + "tpot_p95_ms": 23.99643604938105, + "tpot_p99_ms": 24.379314330662144, + "e2e_p50_ms": 6627.143721503671, + "e2e_p95_ms": 6811.619238646381, + "e2e_p99_ms": 6835.494205219293, + "itl_p50_ms": 15.377331001218408, + "itl_p95_ms": 16.272442803892773, + "itl_p99_ms": 18.94523348833901, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i4096_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i4096_o4096_a1.json new file mode 100644 index 0000000..94b37a8 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i4096_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 316.8896808080026, + "request_tps": 0.252453787059322, + "input_tps": 1034.0507117949828, + "output_tps": 1034.0507117949828, + "total_tps": 2068.1014235899656, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 658.9246715011541, + "ttft_p95_ms": 1005.5294943493209, + "ttft_p99_ms": 1026.091546237294, + "tpot_p50_ms": 15.320106219048359, + "tpot_p95_ms": 15.378604879511569, + "tpot_p99_ms": 15.45278130631852, + "e2e_p50_ms": 63453.26198550174, + "e2e_p95_ms": 63513.736396850436, + "e2e_p99_ms": 63521.74632662762, + "itl_p50_ms": 15.256563012371771, + "itl_p95_ms": 15.777264998177998, + "itl_p99_ms": 16.359419524669644, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i4096_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i4096_o512_a1.json new file mode 100644 index 0000000..66f3453 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i4096_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 48.17216287400515, + "request_tps": 1.6607101534809827, + "input_tps": 6802.268788658105, + "output_tps": 850.2835985822632, + "total_tps": 7652.552387240369, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 1226.4635695100878, + "ttft_p95_ms": 1866.224592289654, + "ttft_p99_ms": 2426.2945817218856, + "tpot_p50_ms": 16.140000863013167, + "tpot_p95_ms": 17.640392132493695, + "tpot_p99_ms": 18.638964293663246, + "e2e_p50_ms": 9504.232979503286, + "e2e_p95_ms": 10157.524074796675, + "e2e_p99_ms": 10178.896629959345, + "itl_p50_ms": 15.288406488252804, + "itl_p95_ms": 15.857895543740598, + "itl_p99_ms": 17.156079869600944, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i65536_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i65536_o1024_a1.json new file mode 100644 index 0000000..d3b9611 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i65536_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 376.9979162859963, + "request_tps": 0.21220276437631763, + "input_tps": 13906.920366166352, + "output_tps": 217.29563072134926, + "total_tps": 14124.215996887702, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 25524.604717495095, + "ttft_p95_ms": 56727.187289898575, + "ttft_p99_ms": 65385.280855649864, + "tpot_p50_ms": 45.977114398337235, + "tpot_p95_ms": 68.72763672496647, + "tpot_p99_ms": 71.65886099602294, + "e2e_p50_ms": 72648.60455600865, + "e2e_p95_ms": 102850.95250014929, + "e2e_p99_ms": 128267.6254233617, + "itl_p50_ms": 15.504026996495668, + "itl_p95_ms": 16.275252601190004, + "itl_p99_ms": 1222.4654011012171, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i65536_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i65536_o128_a1.json new file mode 100644 index 0000000..610bdb1 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i65536_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 263.11688939499436, + "request_tps": 0.30404737675316235, + "input_tps": 19926.048882895248, + "output_tps": 38.91806422440478, + "total_tps": 19964.966947119654, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 21783.06187000271, + "ttft_p95_ms": 48708.6885400975, + "ttft_p99_ms": 53530.08918488397, + "tpot_p50_ms": 208.85715699601343, + "tpot_p95_ms": 353.9625207657282, + "tpot_p99_ms": 368.66045906663794, + "e2e_p50_ms": 53675.07621049299, + "e2e_p95_ms": 64500.535845247214, + "e2e_p99_ms": 78019.4687606334, + "itl_p50_ms": 15.624218496668618, + "itl_p95_ms": 1377.797994745198, + "itl_p99_ms": 1593.279839480674, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i65536_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i65536_o256_a1.json new file mode 100644 index 0000000..e0c732e --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i65536_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 318.6807930810028, + "request_tps": 0.2510348967898592, + "input_tps": 16451.82299602021, + "output_tps": 64.26493357820395, + "total_tps": 16516.087929598416, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 25208.387847495032, + "ttft_p95_ms": 57570.60906205442, + "ttft_p99_ms": 64559.6027293129, + "tpot_p50_ms": 139.20286317058046, + "tpot_p95_ms": 218.68997558943673, + "tpot_p99_ms": 237.28532163909728, + "e2e_p50_ms": 62259.9117714999, + "e2e_p95_ms": 93757.1761358522, + "e2e_p99_ms": 109198.36791741903, + "itl_p50_ms": 15.581318992190063, + "itl_p95_ms": 1060.9013718029018, + "itl_p99_ms": 1579.705025086296, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i65536_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i65536_o512_a1.json new file mode 100644 index 0000000..ecf98cb --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c16_i65536_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 337.8231221129972, + "request_tps": 0.23681031511289238, + "input_tps": 15519.600811238515, + "output_tps": 121.2468813378009, + "total_tps": 15640.847692576317, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 24695.115143498697, + "ttft_p95_ms": 57546.68756374739, + "ttft_p99_ms": 64468.70614748111, + "tpot_p50_ms": 76.04405581409218, + "tpot_p95_ms": 120.77805700313024, + "tpot_p99_ms": 135.3599259530104, + "e2e_p50_ms": 66170.59861750022, + "e2e_p95_ms": 92553.88260725957, + "e2e_p99_ms": 109859.61571042174, + "itl_p50_ms": 15.599313002894633, + "itl_p95_ms": 824.2479222499242, + "itl_p99_ms": 1410.173195308016, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c1_i131072_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c1_i131072_o256_a1.json new file mode 100644 index 0000000..e56a2be --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c1_i131072_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 97.25075286399806, + "request_tps": 0.051413483728936615, + "input_tps": 6738.86813931918, + "output_tps": 13.161851834607774, + "total_tps": 6752.029991153788, + "mean_input_tokens": 131072.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 16969.752281991532, + "ttft_p95_ms": 17121.84501980082, + "ttft_p99_ms": 17125.887859960785, + "tpot_p50_ms": 9.60843350975683, + "tpot_p95_ms": 9.634889066653034, + "tpot_p99_ms": 9.637108200772976, + "e2e_p50_ms": 19427.356342988787, + "e2e_p95_ms": 19571.998772607185, + "e2e_p99_ms": 19575.258900129702, + "itl_p50_ms": 10.00273649697192, + "itl_p95_ms": 10.19823560709483, + "itl_p99_ms": 11.044782063545426, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c1_i131072_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c1_i131072_o512_a1.json new file mode 100644 index 0000000..6039a8f --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c1_i131072_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 110.12223424699914, + "request_tps": 0.04540409150058859, + "input_tps": 5951.205081165148, + "output_tps": 23.24689484830136, + "total_tps": 5974.45197601345, + "mean_input_tokens": 131072.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 16962.474314001156, + "ttft_p95_ms": 17155.541892390465, + "ttft_p99_ms": 17171.51858087047, + "tpot_p50_ms": 9.821654945213728, + "tpot_p95_ms": 9.835613834436566, + "tpot_p99_ms": 9.837956189974383, + "e2e_p50_ms": 21989.969162998023, + "e2e_p95_ms": 22173.37846419541, + "e2e_p99_ms": 22190.17843683483, + "itl_p50_ms": 9.989651007344946, + "itl_p95_ms": 10.151538599166088, + "itl_p99_ms": 10.469720991677605, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c1_i262144_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c1_i262144_o128_a1.json new file mode 100644 index 0000000..ffc3fcb --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c1_i262144_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 215.86107532100868, + "request_tps": 0.02316304591999489, + "input_tps": 6072.05350965114, + "output_tps": 2.964869877759346, + "total_tps": 6075.0183795289, + "mean_input_tokens": 262144.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 42189.58094799018, + "ttft_p95_ms": 42206.01177979843, + "ttft_p99_ms": 42207.67736715847, + "tpot_p50_ms": 9.026367795253979, + "tpot_p95_ms": 9.051597466134083, + "tpot_p99_ms": 9.053308627064942, + "e2e_p50_ms": 43335.3122970002, + "e2e_p95_ms": 43353.944078995846, + "e2e_p99_ms": 43357.12344699539, + "itl_p50_ms": 10.197708994382992, + "itl_p95_ms": 10.398621491185622, + "itl_p99_ms": 11.979914264811663, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c1_i262144_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c1_i262144_o256_a1.json new file mode 100644 index 0000000..631b127 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c1_i262144_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 222.6894240649999, + "request_tps": 0.022452795057481358, + "input_tps": 5885.865507548393, + "output_tps": 5.7479155347152275, + "total_tps": 5891.6134230831085, + "mean_input_tokens": 262144.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 41992.54209799983, + "ttft_p95_ms": 42385.26861999417, + "ttft_p99_ms": 42405.3075527912, + "tpot_p50_ms": 9.564196976419884, + "tpot_p95_ms": 9.778543265853717, + "tpot_p99_ms": 9.797764013144736, + "e2e_p50_ms": 44431.4123269869, + "e2e_p95_ms": 44819.5239008026, + "e2e_p99_ms": 44832.609473761986, + "itl_p50_ms": 10.17798750399379, + "itl_p95_ms": 10.34693365872954, + "itl_p99_ms": 10.991512061009418, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i1024_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i1024_o1024_a1.json new file mode 100644 index 0000000..92b50e3 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i1024_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 99.20981571500306, + "request_tps": 1.6127436468547327, + "input_tps": 1651.4494943792463, + "output_tps": 1651.4494943792463, + "total_tps": 3302.8989887584926, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 1029.0807624987792, + "ttft_p95_ms": 1823.7564543480403, + "ttft_p99_ms": 1851.6841998742893, + "tpot_p50_ms": 18.9562778137816, + "tpot_p95_ms": 19.887580249860058, + "tpot_p99_ms": 19.937121249460077, + "e2e_p50_ms": 20452.989998499106, + "e2e_p95_ms": 20839.25912065315, + "e2e_p99_ms": 20861.500803281087, + "itl_p50_ms": 18.491245500626974, + "itl_p95_ms": 19.586568658996843, + "itl_p99_ms": 20.498206740885504, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i1024_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i1024_o2048_a1.json new file mode 100644 index 0000000..1087187 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i1024_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 197.1833312999952, + "request_tps": 0.8114276138107009, + "input_tps": 830.9018765421578, + "output_tps": 1661.8037530843155, + "total_tps": 2492.705629626473, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 656.5913640006329, + "ttft_p95_ms": 1290.7357379466703, + "ttft_p99_ms": 1304.3246922609978, + "tpot_p50_ms": 18.853423953588848, + "tpot_p95_ms": 19.17381441135775, + "tpot_p99_ms": 19.217155080916477, + "e2e_p50_ms": 39465.7350474954, + "e2e_p95_ms": 39581.901998698595, + "e2e_p99_ms": 39587.70705576069, + "itl_p50_ms": 18.70039399364032, + "itl_p95_ms": 19.59088460134808, + "itl_p99_ms": 20.509281637496308, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i1024_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i1024_o256_a1.json new file mode 100644 index 0000000..e27818f --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i1024_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 58.333280092992936, + "request_tps": 2.742859646241964, + "input_tps": 2808.6882777517712, + "output_tps": 702.1720694379428, + "total_tps": 3510.8603471897136, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 1302.453588003118, + "ttft_p95_ms": 25865.006712951435, + "ttft_p99_ms": 25871.150527912396, + "tpot_p50_ms": 22.315067135302037, + "tpot_p95_ms": 116.87548523570477, + "tpot_p99_ms": 116.9072278467824, + "e2e_p50_ms": 7041.645481498563, + "e2e_p95_ms": 30671.992690498882, + "e2e_p99_ms": 30678.4747339561, + "itl_p50_ms": 18.8346959912451, + "itl_p95_ms": 20.653201197274026, + "itl_p99_ms": 25.224988315603593, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i1024_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i1024_o4096_a1.json new file mode 100644 index 0000000..910132b --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i1024_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 390.9258493729867, + "request_tps": 0.40928477934275004, + "input_tps": 419.10761404697604, + "output_tps": 1676.4304561879042, + "total_tps": 2095.5380702348803, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 1059.6694034975371, + "ttft_p95_ms": 1531.6532846380142, + "ttft_p99_ms": 1558.6441262916196, + "tpot_p50_ms": 18.821365679975997, + "tpot_p95_ms": 19.016937038081775, + "tpot_p99_ms": 19.073814857857453, + "e2e_p50_ms": 77972.91494200181, + "e2e_p95_ms": 78736.66218614817, + "e2e_p99_ms": 78754.35213367702, + "itl_p50_ms": 18.73732249805471, + "itl_p95_ms": 19.672631401772378, + "itl_p99_ms": 20.632040495693218, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i4096_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i4096_o1024_a1.json new file mode 100644 index 0000000..df26fff --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i4096_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 107.34333604600397, + "request_tps": 1.490544321553683, + "input_tps": 6105.269541083886, + "output_tps": 1526.3173852709715, + "total_tps": 7631.5869263548575, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 1697.2974899981637, + "ttft_p95_ms": 2827.3090714494147, + "ttft_p99_ms": 3041.8824867265357, + "tpot_p50_ms": 19.041262989738744, + "tpot_p95_ms": 20.483581416088622, + "tpot_p99_ms": 20.796884204547236, + "e2e_p50_ms": 21322.126428000047, + "e2e_p95_ms": 21977.338839244476, + "e2e_p99_ms": 21997.641324478464, + "itl_p50_ms": 18.680188004509546, + "itl_p95_ms": 19.772065294091586, + "itl_p99_ms": 21.03478136268678, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i4096_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i4096_o2048_a1.json new file mode 100644 index 0000000..e953c0c --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i4096_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 203.61964830600482, + "request_tps": 0.7857787857463927, + "input_tps": 3218.5499064172245, + "output_tps": 1609.2749532086123, + "total_tps": 4827.824859625837, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 1823.4369839992723, + "ttft_p95_ms": 2880.811745747633, + "ttft_p99_ms": 2898.1182565601193, + "tpot_p50_ms": 18.92700290888939, + "tpot_p95_ms": 19.648448953663472, + "tpot_p99_ms": 19.892009308511824, + "e2e_p50_ms": 40591.624035492714, + "e2e_p95_ms": 41278.74980429624, + "e2e_p99_ms": 41301.732580115786, + "itl_p50_ms": 18.753395997919142, + "itl_p95_ms": 19.575604447891237, + "itl_p99_ms": 20.460386879567523, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i4096_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i4096_o256_a1.json new file mode 100644 index 0000000..7bc139f --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i4096_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 66.26290708300075, + "request_tps": 2.414623913188481, + "input_tps": 9890.299548420018, + "output_tps": 618.1437217762511, + "total_tps": 10508.443270196269, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 3032.6789590035332, + "ttft_p95_ms": 5014.513322354456, + "ttft_p99_ms": 14411.20142712994, + "tpot_p50_ms": 25.094734256865696, + "tpot_p95_ms": 108.72940146214187, + "tpot_p99_ms": 110.79764089783636, + "e2e_p50_ms": 9434.836903004907, + "e2e_p95_ms": 29872.09643335227, + "e2e_p99_ms": 29884.739060596476, + "itl_p50_ms": 18.683340997085907, + "itl_p95_ms": 19.744293195253704, + "itl_p99_ms": 26.305865632602966, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i4096_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i4096_o4096_a1.json new file mode 100644 index 0000000..73299cc --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i4096_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 389.95699894400605, + "request_tps": 0.4103016497543987, + "input_tps": 1680.595557394017, + "output_tps": 1680.595557394017, + "total_tps": 3361.191114788034, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 1123.4332419990096, + "ttft_p95_ms": 1676.0268465513946, + "ttft_p99_ms": 1717.2243972464637, + "tpot_p50_ms": 18.85827924127029, + "tpot_p95_ms": 19.032755866286788, + "tpot_p99_ms": 19.08965374972383, + "e2e_p50_ms": 77915.13783499977, + "e2e_p95_ms": 78629.53818804745, + "e2e_p99_ms": 78646.88334376202, + "itl_p50_ms": 18.702409994148184, + "itl_p95_ms": 19.55858354413067, + "itl_p99_ms": 20.443303814536193, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i4096_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i4096_o512_a1.json new file mode 100644 index 0000000..1696527 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c32_i4096_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 63.387327321004705, + "request_tps": 2.52416384728972, + "input_tps": 10338.975118498693, + "output_tps": 1292.3718898123366, + "total_tps": 11631.347008311028, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 2139.6533224979066, + "ttft_p95_ms": 3791.57651143978, + "ttft_p99_ms": 4309.74766367377, + "tpot_p50_ms": 20.01632839431788, + "tpot_p95_ms": 24.519710790119547, + "tpot_p99_ms": 25.129288290261652, + "e2e_p50_ms": 12805.874345496704, + "e2e_p95_ms": 13800.694600059069, + "e2e_p99_ms": 13819.40531841843, + "itl_p50_ms": 18.705550508457236, + "itl_p95_ms": 20.301984495017674, + "itl_p99_ms": 21.65679321915376, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c48_i1024_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c48_i1024_o1024_a1.json new file mode 100644 index 0000000..1594b18 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c48_i1024_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 108.07686458299577, + "request_tps": 2.2206417712617497, + "input_tps": 2273.9371737720317, + "output_tps": 2273.9371737720317, + "total_tps": 4547.874347544063, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 1411.8448265071493, + "ttft_p95_ms": 2348.3333344942366, + "ttft_p99_ms": 2881.5906707521935, + "tpot_p50_ms": 19.57261614369442, + "tpot_p95_ms": 20.474906701757522, + "tpot_p99_ms": 20.999123584376957, + "e2e_p50_ms": 21663.715992996003, + "e2e_p95_ms": 22356.912894507695, + "e2e_p99_ms": 22380.086564826343, + "itl_p50_ms": 18.948694996652193, + "itl_p95_ms": 20.273402400198393, + "itl_p99_ms": 21.75828296574764, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c48_i1024_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c48_i1024_o2048_a1.json new file mode 100644 index 0000000..944a408 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c48_i1024_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 200.68862108800386, + "request_tps": 1.195882450628617, + "input_tps": 1224.5836294437038, + "output_tps": 2449.1672588874076, + "total_tps": 3673.7508883311116, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 833.4129540016875, + "ttft_p95_ms": 1647.5724648524192, + "ttft_p99_ms": 1656.5784132984118, + "tpot_p50_ms": 19.151686465800864, + "tpot_p95_ms": 19.465789479505947, + "tpot_p99_ms": 19.511620899212012, + "e2e_p50_ms": 40142.91101550043, + "e2e_p95_ms": 40502.659226708056, + "e2e_p99_ms": 40515.6316231635, + "itl_p50_ms": 18.947302000015043, + "itl_p95_ms": 20.123915805015713, + "itl_p99_ms": 21.517608713475045, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c48_i1024_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c48_i1024_o4096_a1.json new file mode 100644 index 0000000..d0fe36e --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c48_i1024_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 395.37849693700264, + "request_tps": 0.6070132843826361, + "input_tps": 621.5816032078194, + "output_tps": 2486.3264128312776, + "total_tps": 3107.9080160390968, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 1144.204484502552, + "ttft_p95_ms": 1999.596584997198, + "ttft_p99_ms": 2035.2585943651502, + "tpot_p50_ms": 18.99256894627527, + "tpot_p95_ms": 19.224950862944212, + "tpot_p99_ms": 19.230104005436754, + "e2e_p50_ms": 78907.89274750568, + "e2e_p95_ms": 79839.60575450546, + "e2e_p99_ms": 79862.91444651186, + "itl_p50_ms": 18.858423005440272, + "itl_p95_ms": 20.01599159266334, + "itl_p99_ms": 21.22574228560552, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c48_i4096_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c48_i4096_o1024_a1.json new file mode 100644 index 0000000..93b7d02 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c48_i4096_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 114.72205163900799, + "request_tps": 2.092012795893852, + "input_tps": 8568.884411981218, + "output_tps": 2142.2211029953046, + "total_tps": 10711.105514976523, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 2002.436676499201, + "ttft_p95_ms": 5319.875290898926, + "ttft_p99_ms": 6160.708812107041, + "tpot_p50_ms": 19.824865686698722, + "tpot_p95_ms": 23.0216869557243, + "tpot_p99_ms": 23.20206612341515, + "e2e_p50_ms": 21923.11628299649, + "e2e_p95_ms": 25281.340099497902, + "e2e_p99_ms": 25297.796992440562, + "itl_p50_ms": 18.851815999369137, + "itl_p95_ms": 20.00413840287365, + "itl_p99_ms": 21.528728755656623, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c48_i4096_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c48_i4096_o2048_a1.json new file mode 100644 index 0000000..bc24536 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c48_i4096_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 209.64498886499496, + "request_tps": 1.1447924479346976, + "input_tps": 4689.069866740521, + "output_tps": 2344.5349333702607, + "total_tps": 7033.604800110782, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 1850.1262010031496, + "ttft_p95_ms": 5557.007933600835, + "ttft_p99_ms": 5728.514949967212, + "tpot_p50_ms": 19.032042742548224, + "tpot_p95_ms": 20.8632619918908, + "tpot_p99_ms": 21.040059630580483, + "e2e_p50_ms": 40867.48082550184, + "e2e_p95_ms": 44286.82900869899, + "e2e_p99_ms": 44307.41918329746, + "itl_p50_ms": 18.80523749423446, + "itl_p95_ms": 20.01578960043844, + "itl_p99_ms": 21.193549290037485, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c48_i4096_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c48_i4096_o4096_a1.json new file mode 100644 index 0000000..8584a0d --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c48_i4096_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 402.4469851330068, + "request_tps": 0.596351839785012, + "input_tps": 2442.6571357594094, + "output_tps": 2442.6571357594094, + "total_tps": 4885.314271518819, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 2047.3789034949732, + "ttft_p95_ms": 4081.0472775985545, + "ttft_p99_ms": 4647.203718657256, + "tpot_p50_ms": 19.02283846043863, + "tpot_p95_ms": 19.43320571131982, + "tpot_p99_ms": 19.899421790027564, + "e2e_p50_ms": 80315.7795865045, + "e2e_p95_ms": 81844.74676285972, + "e2e_p99_ms": 81865.64722886324, + "itl_p50_ms": 18.825628008926287, + "itl_p95_ms": 19.975638002506457, + "itl_p99_ms": 21.221037310315303, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c48_i4096_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c48_i4096_o512_a1.json new file mode 100644 index 0000000..657cf85 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c48_i4096_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 77.56397197001206, + "request_tps": 3.0942200857479203, + "input_tps": 12673.925471223482, + "output_tps": 1584.2406839029352, + "total_tps": 14258.166155126417, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 2946.3751199946273, + "ttft_p95_ms": 6820.920404194476, + "ttft_p99_ms": 8050.301623321715, + "tpot_p50_ms": 23.094741272006033, + "tpot_p95_ms": 31.340273439141725, + "tpot_p99_ms": 33.989053702566764, + "e2e_p50_ms": 15109.904168501089, + "e2e_p95_ms": 19630.664032053755, + "e2e_p99_ms": 19649.577008009364, + "itl_p50_ms": 19.291372009320185, + "itl_p95_ms": 22.182088205590844, + "itl_p99_ms": 24.491042752633813, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c64_i1024_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c64_i1024_o1024_a1.json new file mode 100644 index 0000000..526872a --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c64_i1024_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 320, + "failed": 0, + "duration_s": 123.36385090599651, + "request_tps": 2.593952747501702, + "input_tps": 2656.207613441743, + "output_tps": 2656.207613441743, + "total_tps": 5312.415226883486, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 1434.1024970053695, + "ttft_p95_ms": 3039.2205538046255, + "ttft_p99_ms": 3390.209743604064, + "tpot_p50_ms": 22.264150537145387, + "tpot_p95_ms": 23.93735160308324, + "tpot_p99_ms": 24.057023045675336, + "e2e_p50_ms": 24811.180071505078, + "e2e_p95_ms": 25911.275261299306, + "e2e_p99_ms": 25928.425582967902, + "itl_p50_ms": 21.702708501834422, + "itl_p95_ms": 23.22826191157219, + "itl_p99_ms": 25.138318116369188, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c64_i1024_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c64_i1024_o2048_a1.json new file mode 100644 index 0000000..efb8af9 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c64_i1024_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 320, + "failed": 0, + "duration_s": 231.29285078399698, + "request_tps": 1.3835274152024963, + "input_tps": 1416.7320731673562, + "output_tps": 2833.4641463347125, + "total_tps": 4250.196219502069, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 1245.4625594982645, + "ttft_p95_ms": 2417.803988052765, + "ttft_p99_ms": 2683.46135836342, + "tpot_p50_ms": 21.843712060574735, + "tpot_p95_ms": 22.34440391123519, + "tpot_p99_ms": 22.46978759702825, + "e2e_p50_ms": 46248.58569000207, + "e2e_p95_ms": 46951.28065479075, + "e2e_p99_ms": 46979.07613164149, + "itl_p50_ms": 21.61537500069244, + "itl_p95_ms": 23.96329630355467, + "itl_p99_ms": 25.871719614806352, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c64_i1024_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c64_i1024_o4096_a1.json new file mode 100644 index 0000000..a7e2b4a --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c64_i1024_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 320, + "failed": 0, + "duration_s": 443.57241661399894, + "request_tps": 0.7214154623110101, + "input_tps": 738.7294334064743, + "output_tps": 2954.917733625897, + "total_tps": 3693.6471670323713, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 1471.407580494997, + "ttft_p95_ms": 2597.0414660041592, + "ttft_p99_ms": 2976.737847158656, + "tpot_p50_ms": 21.723876868741236, + "tpot_p95_ms": 22.00444482439625, + "tpot_p99_ms": 22.07639913107326, + "e2e_p50_ms": 90742.18811100582, + "e2e_p95_ms": 91541.6523386979, + "e2e_p99_ms": 91559.50254334108, + "itl_p50_ms": 21.4493349994882, + "itl_p95_ms": 22.813913105346728, + "itl_p99_ms": 24.170049769309127, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c80_i1024_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c80_i1024_o1024_a1.json new file mode 100644 index 0000000..103bee8 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c80_i1024_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 400, + "failed": 0, + "duration_s": 143.43459790300403, + "request_tps": 2.7887274468499945, + "input_tps": 2855.6569055743944, + "output_tps": 2855.6569055743944, + "total_tps": 5711.313811148789, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 1577.1677885059034, + "ttft_p95_ms": 5509.251923993023, + "ttft_p99_ms": 20218.74289565443, + "tpot_p50_ms": 22.94427521847951, + "tpot_p95_ms": 40.54245714501829, + "tpot_p99_ms": 40.55054783283226, + "e2e_p50_ms": 24977.686929501942, + "e2e_p95_ms": 42717.079425647535, + "e2e_p99_ms": 42743.186112658295, + "itl_p50_ms": 21.95263899920974, + "itl_p95_ms": 23.614524496224476, + "itl_p99_ms": 25.837959441560088, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c80_i1024_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c80_i1024_o2048_a1.json new file mode 100644 index 0000000..741eea5 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c80_i1024_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 400, + "failed": 0, + "duration_s": 235.45233723000274, + "request_tps": 1.6988576316796469, + "input_tps": 1739.6302148399584, + "output_tps": 3479.260429679917, + "total_tps": 5218.890644519875, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 1586.3785705005284, + "ttft_p95_ms": 3133.5033947099873, + "ttft_p99_ms": 3142.967526005523, + "tpot_p50_ms": 22.12632127210705, + "tpot_p95_ms": 22.91146984701974, + "tpot_p99_ms": 22.976990950927, + "e2e_p50_ms": 47054.46670300444, + "e2e_p95_ms": 48105.06032910416, + "e2e_p99_ms": 48135.96432496226, + "itl_p50_ms": 21.907794500293676, + "itl_p95_ms": 23.54792225087294, + "itl_p99_ms": 25.59503832002519, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c80_i1024_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c80_i1024_o4096_a1.json new file mode 100644 index 0000000..5ebd9ff --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c80_i1024_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 400, + "failed": 0, + "duration_s": 463.4503973050014, + "request_tps": 0.8630912872791345, + "input_tps": 883.8054781738338, + "output_tps": 3535.221912695335, + "total_tps": 4419.027390869169, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 2439.6097005082993, + "ttft_p95_ms": 3686.9959791976726, + "ttft_p99_ms": 3709.631283667256, + "tpot_p50_ms": 22.095354364345052, + "tpot_p95_ms": 22.518257637104984, + "tpot_p99_ms": 22.625587380977006, + "e2e_p50_ms": 92209.49949500209, + "e2e_p95_ms": 93844.80823813865, + "e2e_p99_ms": 93874.67774137796, + "itl_p50_ms": 21.87070349464193, + "itl_p95_ms": 23.618548751983326, + "itl_p99_ms": 25.773182249395177, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c8_i131072_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c8_i131072_o256_a1.json new file mode 100644 index 0000000..992af65 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c8_i131072_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 341.0657499819936, + "request_tps": 0.11727943952775025, + "input_tps": 15372.05069778128, + "output_tps": 30.023536519104063, + "total_tps": 15402.074234300384, + "mean_input_tokens": 131072.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 38767.44465551019, + "ttft_p95_ms": 67753.44985890988, + "ttft_p99_ms": 68613.70976785096, + "tpot_p50_ms": 99.49740432157162, + "tpot_p95_ms": 179.5039385709941, + "tpot_p99_ms": 227.61194875988735, + "e2e_p50_ms": 70759.12754750607, + "e2e_p95_ms": 71843.55354870277, + "e2e_p99_ms": 71887.88047623617, + "itl_p50_ms": 12.980383995454758, + "itl_p95_ms": 1125.97120024584, + "itl_p99_ms": 1602.8499752024072, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c8_i131072_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c8_i131072_o512_a1.json new file mode 100644 index 0000000..58cd809 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c8_i131072_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 368.12840555600997, + "request_tps": 0.10865773843120097, + "input_tps": 14241.987091654373, + "output_tps": 55.632762076774895, + "total_tps": 14297.619853731148, + "mean_input_tokens": 131072.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 36897.75435599586, + "ttft_p95_ms": 68026.47409820274, + "ttft_p99_ms": 74667.961793141, + "tpot_p50_ms": 63.40141743053057, + "tpot_p95_ms": 116.81624726055898, + "tpot_p99_ms": 143.86918745584376, + "e2e_p50_ms": 74333.1425399956, + "e2e_p95_ms": 83933.02225654986, + "e2e_p99_ms": 104491.18695786949, + "itl_p50_ms": 12.98504300211789, + "itl_p95_ms": 18.927634652209232, + "itl_p99_ms": 1447.1019217558205, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c8_i262144_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c8_i262144_o128_a1.json new file mode 100644 index 0000000..c3d1cd4 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c8_i262144_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 1143.3382095730049, + "request_tps": 0.03498527353068918, + "input_tps": 9171.179544428984, + "output_tps": 4.478115011928215, + "total_tps": 9175.65765944091, + "mean_input_tokens": 262144.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 60434.31567850348, + "ttft_p95_ms": 171402.15240599928, + "ttft_p99_ms": 221100.73346630885, + "tpot_p50_ms": 1432.7522763937384, + "tpot_p95_ms": 1566.4298780566398, + "tpot_p99_ms": 1577.578224665822, + "e2e_p50_ms": 243043.1625314959, + "e2e_p95_ms": 344139.9401556495, + "e2e_p99_ms": 419677.4824362584, + "itl_p50_ms": 1394.7447990067303, + "itl_p95_ms": 1954.1919693932869, + "itl_p99_ms": 2118.3612401573937, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c8_i262144_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c8_i262144_o256_a1.json new file mode 100644 index 0000000..91ead40 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c8_i262144_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 818.7336148600007, + "request_tps": 0.04885593955591991, + "input_tps": 12807.291418947068, + "output_tps": 12.507120526315497, + "total_tps": 12819.798539473382, + "mean_input_tokens": 262144.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 105032.74306450476, + "ttft_p95_ms": 162785.75602490891, + "ttft_p99_ms": 165083.89961342298, + "tpot_p50_ms": 214.14210347058685, + "tpot_p95_ms": 422.65357399257476, + "tpot_p99_ms": 440.83721461195177, + "e2e_p50_ms": 164185.14075000712, + "e2e_p95_ms": 168861.28058029353, + "e2e_p99_ms": 169246.668313704, + "itl_p50_ms": 13.376075992709957, + "itl_p95_ms": 1720.7928034986253, + "itl_p99_ms": 2045.3944211971263, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c96_i1024_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c96_i1024_o2048_a1.json new file mode 100644 index 0000000..9f5a88b --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c96_i1024_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 480, + "failed": 0, + "duration_s": 261.48785026300175, + "request_tps": 1.8356493409434549, + "input_tps": 1879.7049251260978, + "output_tps": 3759.4098502521956, + "total_tps": 5639.114775378293, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 2183.7680815006024, + "ttft_p95_ms": 3805.7011090502783, + "ttft_p99_ms": 4558.676817088854, + "tpot_p50_ms": 24.876903371278917, + "tpot_p95_ms": 25.891116515634604, + "tpot_p99_ms": 26.181815688893472, + "e2e_p50_ms": 53079.94003850763, + "e2e_p95_ms": 54830.756290954014, + "e2e_p99_ms": 54843.78045572128, + "itl_p50_ms": 24.47853601188399, + "itl_p95_ms": 26.07234639581293, + "itl_p99_ms": 27.980867308797315, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c96_i1024_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c96_i1024_o4096_a1.json new file mode 100644 index 0000000..4e0c207 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/metrics/vllm_c96_i1024_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 480, + "failed": 0, + "duration_s": 524.4931549339963, + "request_tps": 0.9151692362132059, + "input_tps": 937.1332978823228, + "output_tps": 3748.5331915292913, + "total_tps": 4685.666489411614, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 2751.5591720002703, + "ttft_p95_ms": 4227.864060657878, + "ttft_p99_ms": 4536.041375702771, + "tpot_p50_ms": 24.893937553724363, + "tpot_p95_ms": 25.43370572488336, + "tpot_p99_ms": 25.74416057706036, + "e2e_p50_ms": 104575.3224170112, + "e2e_p95_ms": 106446.31684675842, + "e2e_p99_ms": 106478.32935052022, + "itl_p50_ms": 24.56568699562922, + "itl_p95_ms": 27.559315803227946, + "itl_p99_ms": 30.286136486683972, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/server_cmd.txt b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/server_cmd.txt new file mode 100644 index 0000000..0fd28a6 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp2_dp4/server_cmd.txt @@ -0,0 +1 @@ +vllm serve /data1/hf_models/DeepSeek-V4-Flash --trust-remote-code --kv-cache-dtype fp8 --block-size 256 --tensor-parallel-size 2 --gpu-memory-utilization 0.9 --max-model-len 1048576 --max-num-seqs 128 --no-enable-flashinfer-autotune --host 0.0.0.0 --port 30030 --data-parallel-size 4 diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c112_i1024_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c112_i1024_o1024_a1.json new file mode 100644 index 0000000..f5b52f7 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c112_i1024_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 560, + "failed": 0, + "duration_s": 143.62631484199665, + "request_tps": 3.899006951588615, + "input_tps": 3992.5831184267417, + "output_tps": 3992.5831184267417, + "total_tps": 7985.1662368534835, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 1571.4164454911952, + "ttft_p95_ms": 4123.8772861033785, + "ttft_p99_ms": 4936.747990879667, + "tpot_p50_ms": 25.948950879762915, + "tpot_p95_ms": 28.919239563878318, + "tpot_p99_ms": 29.696389907960754, + "e2e_p50_ms": 28112.99813849473, + "e2e_p95_ms": 31451.82638790211, + "e2e_p99_ms": 31470.12428371294, + "itl_p50_ms": 25.19894449505955, + "itl_p95_ms": 28.669426793931052, + "itl_p99_ms": 31.52743758109864, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i1024_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i1024_o1024_a1.json new file mode 100644 index 0000000..2a947ff --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i1024_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 70.41322445700644, + "request_tps": 1.1361502134992716, + "input_tps": 1163.417818623254, + "output_tps": 1163.417818623254, + "total_tps": 2326.835637246508, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 421.1962485060212, + "ttft_p95_ms": 677.2944091026147, + "ttft_p99_ms": 691.2584483872342, + "tpot_p50_ms": 12.164721241445186, + "tpot_p95_ms": 15.51125149233277, + "tpot_p99_ms": 15.721278570422475, + "e2e_p50_ms": 12909.268272502231, + "e2e_p95_ms": 16315.028272505877, + "e2e_p99_ms": 16321.569995415775, + "itl_p50_ms": 12.13237400224898, + "itl_p95_ms": 15.468433196656404, + "itl_p99_ms": 16.074650280061178, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i1024_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i1024_o128_a1.json new file mode 100644 index 0000000..6358aeb --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i1024_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 52.56799242499983, + "request_tps": 1.5218386000594972, + "input_tps": 1558.3627264609252, + "output_tps": 194.79534080761564, + "total_tps": 1753.158067268541, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 937.427844000922, + "ttft_p95_ms": 1428.1466051499592, + "ttft_p99_ms": 18393.34938969926, + "tpot_p50_ms": 19.7043993346339, + "tpot_p95_ms": 181.49350103310894, + "tpot_p99_ms": 182.55337019226835, + "e2e_p50_ms": 3306.429382006172, + "e2e_p95_ms": 23952.145700105757, + "e2e_p99_ms": 23955.10612057842, + "itl_p50_ms": 14.576606998161878, + "itl_p95_ms": 15.284183846233645, + "itl_p99_ms": 20.36303154163767, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i1024_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i1024_o2048_a1.json new file mode 100644 index 0000000..6b3abaf --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i1024_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 137.77100664599857, + "request_tps": 0.5806736986800084, + "input_tps": 594.6098674483286, + "output_tps": 1189.2197348966572, + "total_tps": 1783.8296023449857, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 509.03518099948997, + "ttft_p95_ms": 817.3343080510676, + "ttft_p99_ms": 821.0908090611338, + "tpot_p50_ms": 12.053855090129739, + "tpot_p95_ms": 15.30965305794184, + "tpot_p99_ms": 15.332859359112767, + "e2e_p50_ms": 25319.949440003256, + "e2e_p95_ms": 31725.24674650267, + "e2e_p99_ms": 31736.754303508642, + "itl_p50_ms": 12.262283504242077, + "itl_p95_ms": 15.428829462325668, + "itl_p99_ms": 15.964222954062278, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i1024_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i1024_o256_a1.json new file mode 100644 index 0000000..4b581d7 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i1024_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 48.607205292995786, + "request_tps": 1.6458465266162476, + "input_tps": 1685.3468432550376, + "output_tps": 421.3367108137594, + "total_tps": 2106.683554068797, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 782.0743444972322, + "ttft_p95_ms": 12196.891359947767, + "ttft_p99_ms": 12266.554353885149, + "tpot_p50_ms": 15.961954474515354, + "tpot_p95_ms": 58.48537955195449, + "tpot_p99_ms": 60.24822486883116, + "e2e_p50_ms": 8633.398842990573, + "e2e_p95_ms": 16283.070811999643, + "e2e_p99_ms": 16288.124511987555, + "itl_p50_ms": 14.68896050209878, + "itl_p95_ms": 15.627986006438732, + "itl_p99_ms": 16.497610333171902, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i1024_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i1024_o4096_a1.json new file mode 100644 index 0000000..4a790b0 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i1024_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 271.16081980899617, + "request_tps": 0.29502787333491415, + "input_tps": 302.1085422949521, + "output_tps": 1208.4341691798084, + "total_tps": 1510.5427114747604, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 627.1758095026598, + "ttft_p95_ms": 960.0995247899846, + "ttft_p99_ms": 971.8106476443063, + "tpot_p50_ms": 11.913466125030766, + "tpot_p95_ms": 15.119212534556272, + "tpot_p99_ms": 15.134965707044527, + "e2e_p50_ms": 49374.50365599943, + "e2e_p95_ms": 62329.371708749386, + "e2e_p99_ms": 62340.975158869114, + "itl_p50_ms": 12.052630001562648, + "itl_p95_ms": 15.1747610012535, + "itl_p99_ms": 15.672232958022514, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i1024_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i1024_o512_a1.json new file mode 100644 index 0000000..e41faec --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i1024_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 37.51114835499902, + "request_tps": 2.132699304294655, + "input_tps": 2183.8840875977266, + "output_tps": 1091.9420437988633, + "total_tps": 3275.82613139659, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 565.4102255066391, + "ttft_p95_ms": 804.2641735068173, + "ttft_p99_ms": 840.2388689920183, + "tpot_p50_ms": 12.695262536184165, + "tpot_p95_ms": 16.153524499127695, + "tpot_p99_ms": 16.279702295489752, + "e2e_p50_ms": 6936.251829996763, + "e2e_p95_ms": 8593.082947545918, + "e2e_p99_ms": 8608.418859792146, + "itl_p50_ms": 12.14090500434395, + "itl_p95_ms": 15.356732595682843, + "itl_p99_ms": 15.861063279880907, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i131072_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i131072_o128_a1.json new file mode 100644 index 0000000..e9fd589 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i131072_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 782.5810133770137, + "request_tps": 0.10222583813371851, + "input_tps": 13398.945055862752, + "output_tps": 13.084907281115969, + "total_tps": 13412.029963143868, + "mean_input_tokens": 131072.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 58004.37130450155, + "ttft_p95_ms": 110342.86055744666, + "ttft_p99_ms": 145009.39971607734, + "tpot_p50_ms": 726.4293738111247, + "tpot_p95_ms": 1148.8756013172838, + "tpot_p99_ms": 1185.2487841417542, + "e2e_p50_ms": 145932.8185465056, + "e2e_p95_ms": 251712.73757864692, + "e2e_p99_ms": 293697.5071116265, + "itl_p50_ms": 1063.0883960038773, + "itl_p95_ms": 1341.7487418533708, + "itl_p99_ms": 1398.4825665256358, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i16384_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i16384_o1024_a1.json new file mode 100644 index 0000000..bb66243 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i16384_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 81.79191887499474, + "request_tps": 0.9780917369387875, + "input_tps": 16025.055018005094, + "output_tps": 1001.5659386253184, + "total_tps": 17026.620956630413, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 999.8386475053849, + "ttft_p95_ms": 3296.2886180946934, + "ttft_p99_ms": 4053.4733163932106, + "tpot_p50_ms": 15.505134818183233, + "tpot_p95_ms": 18.54402336778967, + "tpot_p99_ms": 18.632150808406323, + "e2e_p50_ms": 16877.44059450779, + "e2e_p95_ms": 19530.405244355643, + "e2e_p99_ms": 19544.808635528025, + "itl_p50_ms": 15.109817002667114, + "itl_p95_ms": 16.077759441395756, + "itl_p99_ms": 16.771573076403005, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i16384_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i16384_o128_a1.json new file mode 100644 index 0000000..b6d315e --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i16384_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 61.62385746999644, + "request_tps": 1.2981985108438008, + "input_tps": 21269.684401664832, + "output_tps": 166.1694093880065, + "total_tps": 21435.85381105284, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 6200.594059002469, + "ttft_p95_ms": 9982.714110245432, + "ttft_p99_ms": 12083.678980552644, + "tpot_p50_ms": 46.56020286221444, + "tpot_p95_ms": 83.96291618064566, + "tpot_p99_ms": 94.82308160958443, + "e2e_p50_ms": 11807.360419501492, + "e2e_p95_ms": 14518.52319239406, + "e2e_p99_ms": 15406.82603497858, + "itl_p50_ms": 14.998138998635113, + "itl_p95_ms": 19.290680799167607, + "itl_p99_ms": 913.8215772807598, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i16384_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i16384_o2048_a1.json new file mode 100644 index 0000000..0c26de3 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i16384_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 136.13383251799678, + "request_tps": 0.5876570028205448, + "input_tps": 9628.172334211806, + "output_tps": 1203.5215417764757, + "total_tps": 10831.693875988281, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 952.042529002938, + "ttft_p95_ms": 1976.2296090004384, + "ttft_p99_ms": 2348.7864647468086, + "tpot_p50_ms": 12.17877617391463, + "tpot_p95_ms": 15.498852642595702, + "tpot_p99_ms": 15.918200653907457, + "e2e_p50_ms": 25627.524990501115, + "e2e_p95_ms": 32975.65399566083, + "e2e_p99_ms": 32987.294595711865, + "itl_p50_ms": 11.984888005827088, + "itl_p95_ms": 15.32018614525441, + "itl_p99_ms": 15.952927653561346, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i16384_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i16384_o256_a1.json new file mode 100644 index 0000000..c5b79c8 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i16384_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 48.023830674996134, + "request_tps": 1.665839623277958, + "input_tps": 27293.116387786064, + "output_tps": 426.45494355915724, + "total_tps": 27719.57133134522, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 4218.072452502383, + "ttft_p95_ms": 7202.70946295495, + "ttft_p99_ms": 9814.834247549588, + "tpot_p50_ms": 20.867704264716902, + "tpot_p95_ms": 43.5013626262719, + "tpot_p99_ms": 48.00088939475241, + "e2e_p50_ms": 10424.10009400919, + "e2e_p95_ms": 13663.989264789414, + "e2e_p99_ms": 16127.52558444379, + "itl_p50_ms": 14.975312995375134, + "itl_p95_ms": 15.900831998442298, + "itl_p99_ms": 664.3842326034808, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i16384_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i16384_o512_a1.json new file mode 100644 index 0000000..72edd0c --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i16384_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 56.001584703000844, + "request_tps": 1.4285310036184244, + "input_tps": 23405.051963284266, + "output_tps": 731.4078738526333, + "total_tps": 24136.4598371369, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 2150.0688629894285, + "ttft_p95_ms": 4893.693922802049, + "ttft_p99_ms": 5737.629971327547, + "tpot_p50_ms": 16.641745551845485, + "tpot_p95_ms": 22.60100227660779, + "tpot_p99_ms": 25.976594107668582, + "e2e_p50_ms": 10941.78693700087, + "e2e_p95_ms": 14099.559878905711, + "e2e_p99_ms": 14118.464289340336, + "itl_p50_ms": 15.120271986234002, + "itl_p95_ms": 16.094310994958505, + "itl_p99_ms": 19.090893008979037, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i4096_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i4096_o1024_a1.json new file mode 100644 index 0000000..80ea27b --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i4096_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 74.54846192999685, + "request_tps": 1.0731274385663692, + "input_tps": 4395.529988367848, + "output_tps": 1098.882497091962, + "total_tps": 5494.41248545981, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 664.4156365000526, + "ttft_p95_ms": 974.0240385996003, + "ttft_p99_ms": 982.1483760114643, + "tpot_p50_ms": 14.938762804988556, + "tpot_p95_ms": 15.543498240374383, + "tpot_p99_ms": 15.80483324938253, + "e2e_p50_ms": 16048.860944996704, + "e2e_p95_ms": 16527.93401440722, + "e2e_p99_ms": 16534.28529751647, + "itl_p50_ms": 14.674316000309773, + "itl_p95_ms": 15.759427801822312, + "itl_p99_ms": 16.294228279730305, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i4096_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i4096_o128_a1.json new file mode 100644 index 0000000..aea7cd4 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i4096_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 63.706278053999995, + "request_tps": 1.2557632064486453, + "input_tps": 5143.606093613651, + "output_tps": 160.7376904254266, + "total_tps": 5304.343784039078, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 2722.8040709960624, + "ttft_p95_ms": 26840.74661414561, + "ttft_p99_ms": 27275.76444137885, + "tpot_p50_ms": 20.65726973231667, + "tpot_p95_ms": 151.34337404136696, + "tpot_p99_ms": 207.36058258637357, + "e2e_p50_ms": 5354.688991996227, + "e2e_p95_ms": 29170.501739544852, + "e2e_p99_ms": 29177.230598040624, + "itl_p50_ms": 14.814941496297251, + "itl_p95_ms": 15.711286505393218, + "itl_p99_ms": 182.99884561245784, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i4096_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i4096_o2048_a1.json new file mode 100644 index 0000000..565a122 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i4096_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 138.75295940799697, + "request_tps": 0.5765642789986448, + "input_tps": 2361.607286778449, + "output_tps": 1180.8036433892246, + "total_tps": 3542.4109301676735, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 632.4307624963694, + "ttft_p95_ms": 959.6396220949828, + "ttft_p99_ms": 969.5783248801308, + "tpot_p50_ms": 12.087159364439152, + "tpot_p95_ms": 15.38081124682696, + "tpot_p99_ms": 15.53437136937437, + "e2e_p50_ms": 25168.70826199738, + "e2e_p95_ms": 32143.357299292988, + "e2e_p99_ms": 32148.356982687113, + "itl_p50_ms": 12.129640497732908, + "itl_p95_ms": 15.666113744373433, + "itl_p99_ms": 16.459487642714517, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i4096_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i4096_o256_a1.json new file mode 100644 index 0000000..3cbe33f --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i4096_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 25.900737181000295, + "request_tps": 3.088715175979033, + "input_tps": 12651.377360810118, + "output_tps": 790.7110850506324, + "total_tps": 13442.08844586075, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 1070.6434219973744, + "ttft_p95_ms": 2601.758133142721, + "ttft_p99_ms": 2731.048192037851, + "tpot_p50_ms": 14.66723954899684, + "tpot_p95_ms": 20.23464126315529, + "tpot_p99_ms": 21.929446337492916, + "e2e_p50_ms": 5262.183128004835, + "e2e_p95_ms": 6524.885636250838, + "e2e_p99_ms": 6530.703649207862, + "itl_p50_ms": 12.11705449532019, + "itl_p95_ms": 15.253629755170547, + "itl_p99_ms": 16.63169343562916, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i4096_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i4096_o4096_a1.json new file mode 100644 index 0000000..03f1a35 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i4096_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 274.1387693459983, + "request_tps": 0.2918230069787383, + "input_tps": 1195.307036584912, + "output_tps": 1195.307036584912, + "total_tps": 2390.614073169824, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 1043.4461850018124, + "ttft_p95_ms": 1874.445231048594, + "ttft_p99_ms": 1897.9588461619276, + "tpot_p50_ms": 11.862804140170747, + "tpot_p95_ms": 15.333563006644244, + "tpot_p99_ms": 15.394116480524199, + "e2e_p50_ms": 49251.109377997636, + "e2e_p95_ms": 63309.41746450044, + "e2e_p99_ms": 63314.34459721451, + "itl_p50_ms": 11.962887001573108, + "itl_p95_ms": 15.223909802443814, + "itl_p99_ms": 15.640087766805664, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i4096_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i4096_o512_a1.json new file mode 100644 index 0000000..fc97182 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i4096_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 37.81364577100612, + "request_tps": 2.1156383725723837, + "input_tps": 8665.654774056484, + "output_tps": 1083.2068467570605, + "total_tps": 9748.861620813545, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 626.4680629974464, + "ttft_p95_ms": 1079.5922750941827, + "ttft_p99_ms": 1085.4407174367225, + "tpot_p50_ms": 12.883891402167945, + "tpot_p95_ms": 15.779685464180407, + "tpot_p99_ms": 16.333569430081702, + "e2e_p50_ms": 7039.921916504682, + "e2e_p95_ms": 8655.18560165001, + "e2e_p99_ms": 8659.915602506953, + "itl_p50_ms": 12.184623003122397, + "itl_p95_ms": 15.375507500721142, + "itl_p99_ms": 16.372877460671585, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i65536_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i65536_o1024_a1.json new file mode 100644 index 0000000..56b3dfc --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i65536_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 390.9239437669894, + "request_tps": 0.2046433872254294, + "input_tps": 13411.509025205742, + "output_tps": 209.5548285188397, + "total_tps": 13621.063853724581, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 34109.15530550119, + "ttft_p95_ms": 49910.73756714177, + "ttft_p99_ms": 58476.69062248939, + "tpot_p50_ms": 42.09774688367832, + "tpot_p95_ms": 69.21228409907829, + "tpot_p99_ms": 69.88106847383611, + "e2e_p50_ms": 79284.3169455009, + "e2e_p95_ms": 90811.3658934919, + "e2e_p99_ms": 100775.45445886119, + "itl_p50_ms": 12.344260001555085, + "itl_p95_ms": 13.471613003639504, + "itl_p99_ms": 1075.7671469193883, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i65536_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i65536_o128_a1.json new file mode 100644 index 0000000..3393c35 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i65536_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 332.67608038699836, + "request_tps": 0.24047415704470515, + "input_tps": 15759.714356081797, + "output_tps": 30.78069210172226, + "total_tps": 15790.495048183519, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 31602.030366491817, + "ttft_p95_ms": 50072.25141515082, + "ttft_p99_ms": 58527.80881931147, + "tpot_p50_ms": 269.0473995314715, + "tpot_p95_ms": 453.42705731063467, + "tpot_p99_ms": 478.3763391067141, + "e2e_p50_ms": 67748.1715354952, + "e2e_p95_ms": 83297.61391548891, + "e2e_p99_ms": 92202.8402509741, + "itl_p50_ms": 12.720453007204924, + "itl_p95_ms": 1088.2944416043756, + "itl_p99_ms": 1121.3325642509153, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i65536_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i65536_o256_a1.json new file mode 100644 index 0000000..513eace --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i65536_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 340.1405706650112, + "request_tps": 0.23519687711345766, + "input_tps": 15413.862538507561, + "output_tps": 60.21040054104516, + "total_tps": 15474.072939048607, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 33961.24007800245, + "ttft_p95_ms": 49937.02013984801, + "ttft_p99_ms": 62166.142975594805, + "tpot_p50_ms": 139.35726555489464, + "tpot_p95_ms": 236.3478919172656, + "tpot_p99_ms": 244.71531864295898, + "e2e_p50_ms": 69747.46793750091, + "e2e_p95_ms": 78453.58004349907, + "e2e_p99_ms": 96399.40396136702, + "itl_p50_ms": 12.44720199611038, + "itl_p95_ms": 1048.4594318972086, + "itl_p99_ms": 1099.1333858296275, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i65536_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i65536_o512_a1.json new file mode 100644 index 0000000..7cfb3f2 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c16_i65536_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 365.2319012020016, + "request_tps": 0.21903891674499099, + "input_tps": 14354.93444779973, + "output_tps": 112.14792537343538, + "total_tps": 14467.082373173163, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 33553.05798899644, + "ttft_p95_ms": 49813.80844564191, + "ttft_p99_ms": 59237.044557964655, + "tpot_p50_ms": 79.3469260440206, + "tpot_p95_ms": 126.30341421105851, + "tpot_p99_ms": 137.98316977467067, + "e2e_p50_ms": 73173.0037334928, + "e2e_p95_ms": 90507.30776954733, + "e2e_p99_ms": 102572.02681419696, + "itl_p50_ms": 12.42508299765177, + "itl_p95_ms": 974.080784598482, + "itl_p99_ms": 1092.6132275909183, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c1_i131072_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c1_i131072_o128_a1.json new file mode 100644 index 0000000..2901a25 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c1_i131072_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 82.18055039300816, + "request_tps": 0.06084164654639981, + "input_tps": 7974.636296129715, + "output_tps": 7.787730757939175, + "total_tps": 7982.424026887654, + "mean_input_tokens": 131072.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 15437.700795999262, + "ttft_p95_ms": 15571.00856200559, + "ttft_p99_ms": 15593.639437204693, + "tpot_p50_ms": 7.764350944867827, + "tpot_p95_ms": 7.770626307061807, + "tpot_p99_ms": 7.771325803097836, + "e2e_p50_ms": 16424.126157013234, + "e2e_p95_ms": 16554.536877613282, + "e2e_p99_ms": 16576.65890193428, + "itl_p50_ms": 8.529996004654095, + "itl_p95_ms": 8.682068198686466, + "itl_p99_ms": 9.645803358289447, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c1_i65536_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c1_i65536_o1024_a1.json new file mode 100644 index 0000000..89d3e32 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c1_i65536_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 76.87229951999325, + "request_tps": 0.06504293524743045, + "input_tps": 4262.653804375602, + "output_tps": 66.60396569336878, + "total_tps": 4329.257770068971, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 6793.178814987186, + "ttft_p95_ms": 6830.732189002447, + "ttft_p99_ms": 6836.927511403337, + "tpot_p50_ms": 8.389067413488618, + "tpot_p95_ms": 8.404996331775118, + "tpot_p99_ms": 8.405754934785953, + "e2e_p50_ms": 15372.661996996612, + "e2e_p95_ms": 15412.706287001492, + "e2e_p99_ms": 15418.935102202231, + "itl_p50_ms": 8.452879003016278, + "itl_p95_ms": 8.561783994082361, + "itl_p99_ms": 8.750601002248002, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c1_i65536_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c1_i65536_o128_a1.json new file mode 100644 index 0000000..e5c1a31 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c1_i65536_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 38.88842260101228, + "request_tps": 0.12857297019472957, + "input_tps": 8426.158174681797, + "output_tps": 16.457340184925386, + "total_tps": 8442.615514866722, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 6736.3390589889605, + "ttft_p95_ms": 6827.516746404581, + "ttft_p99_ms": 6834.052311684936, + "tpot_p50_ms": 7.981791047220794, + "tpot_p95_ms": 8.017796259758422, + "tpot_p99_ms": 8.0200000455801, + "e2e_p50_ms": 7747.252895991551, + "e2e_p95_ms": 7842.1887119999155, + "e2e_p99_ms": 7847.936675201636, + "itl_p50_ms": 8.454770999378525, + "itl_p95_ms": 8.66287379467394, + "itl_p99_ms": 10.578372052987115, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c1_i65536_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c1_i65536_o256_a1.json new file mode 100644 index 0000000..ec541c8 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c1_i65536_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 44.46872818299744, + "request_tps": 0.11243856535370274, + "input_tps": 7368.773819020263, + "output_tps": 28.7842727305479, + "total_tps": 7397.5580917508105, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 6767.2624370025005, + "ttft_p95_ms": 6892.302849600674, + "ttft_p99_ms": 6911.403660320793, + "tpot_p50_ms": 8.220937709819397, + "tpot_p95_ms": 8.2882254510004, + "tpot_p99_ms": 8.295880123954632, + "e2e_p50_ms": 8871.000218990957, + "e2e_p95_ms": 9004.320606408874, + "e2e_p99_ms": 9026.557145289844, + "itl_p50_ms": 8.483409998007119, + "itl_p95_ms": 8.675873003085144, + "itl_p99_ms": 9.017845205380567, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c1_i65536_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c1_i65536_o512_a1.json new file mode 100644 index 0000000..337d8da --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c1_i65536_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 55.36560429999372, + "request_tps": 0.09030877677967596, + "input_tps": 5918.4759950328435, + "output_tps": 46.23809371119409, + "total_tps": 5964.714088744037, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 6785.467125999276, + "ttft_p95_ms": 6867.230469404603, + "ttft_p99_ms": 6880.844480285305, + "tpot_p50_ms": 8.349586817987925, + "tpot_p95_ms": 8.493367793736068, + "tpot_p99_ms": 8.520835669118448, + "e2e_p50_ms": 11042.787049998879, + "e2e_p95_ms": 11206.68308060267, + "e2e_p99_ms": 11234.859840924619, + "itl_p50_ms": 8.474283000396099, + "itl_p95_ms": 8.707773692003684, + "itl_p99_ms": 9.049925783474462, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i1024_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i1024_o1024_a1.json new file mode 100644 index 0000000..e47b46d --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i1024_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 85.3312213080062, + "request_tps": 1.875046407955115, + "input_tps": 1920.0475217460378, + "output_tps": 1920.0475217460378, + "total_tps": 3840.0950434920755, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 806.487787005608, + "ttft_p95_ms": 900.6144404585939, + "ttft_p99_ms": 904.4747244215978, + "tpot_p50_ms": 15.241720033234294, + "tpot_p95_ms": 17.703688128979007, + "tpot_p99_ms": 17.972658989592805, + "e2e_p50_ms": 16181.125543495, + "e2e_p95_ms": 18731.30079265975, + "e2e_p99_ms": 18746.509167428594, + "itl_p50_ms": 15.200200999970548, + "itl_p95_ms": 17.96690479677636, + "itl_p99_ms": 18.968004560447298, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i1024_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i1024_o128_a1.json new file mode 100644 index 0000000..b352162 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i1024_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 95.59839020299842, + "request_tps": 1.6736683500658116, + "input_tps": 1713.836390467391, + "output_tps": 214.22954880842389, + "total_tps": 1928.0659392758148, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 4119.889060501009, + "ttft_p95_ms": 26025.381604798167, + "ttft_p99_ms": 26478.019099681696, + "tpot_p50_ms": 21.413118271642976, + "tpot_p95_ms": 376.97693987998275, + "tpot_p99_ms": 387.07957727625, + "e2e_p50_ms": 7732.499119003478, + "e2e_p95_ms": 51129.2892126934, + "e2e_p99_ms": 51135.8539534836, + "itl_p50_ms": 16.923131996009033, + "itl_p95_ms": 18.388393107306907, + "itl_p99_ms": 244.194259904325, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i1024_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i1024_o2048_a1.json new file mode 100644 index 0000000..c82ff70 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i1024_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 162.71412774299097, + "request_tps": 0.9833196552712499, + "input_tps": 1006.9193269977599, + "output_tps": 2013.8386539955197, + "total_tps": 3020.75798099328, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 871.0656835028203, + "ttft_p95_ms": 1184.4237971468829, + "ttft_p99_ms": 1192.6747960213106, + "tpot_p50_ms": 15.119994677330888, + "tpot_p95_ms": 17.311258725843935, + "tpot_p99_ms": 17.470737866078203, + "e2e_p50_ms": 31956.353274501453, + "e2e_p95_ms": 36126.69280984701, + "e2e_p99_ms": 36139.65432776953, + "itl_p50_ms": 14.982249005697668, + "itl_p95_ms": 17.40834979864303, + "itl_p99_ms": 18.12914603331592, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i1024_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i1024_o4096_a1.json new file mode 100644 index 0000000..f475c20 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i1024_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 318.9912624319986, + "request_tps": 0.50158113667489, + "input_tps": 513.6190839550874, + "output_tps": 2054.4763358203495, + "total_tps": 2568.0954197754368, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 867.0998999950825, + "ttft_p95_ms": 1568.6099555969122, + "ttft_p99_ms": 1761.4919271026156, + "tpot_p50_ms": 14.940362645543985, + "tpot_p95_ms": 17.25645533934084, + "tpot_p99_ms": 17.34749941946349, + "e2e_p50_ms": 62241.984514999785, + "e2e_p95_ms": 71418.52431923835, + "e2e_p99_ms": 71433.78462005581, + "itl_p50_ms": 14.883934498357121, + "itl_p95_ms": 17.468459051451646, + "itl_p99_ms": 18.707020131550955, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i1024_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i1024_o512_a1.json new file mode 100644 index 0000000..73d27d4 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i1024_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 44.825771888004965, + "request_tps": 3.569375233509694, + "input_tps": 3655.040239113927, + "output_tps": 1827.5201195569634, + "total_tps": 5482.56035867089, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 912.5870395073434, + "ttft_p95_ms": 1430.6775802040647, + "ttft_p99_ms": 1454.0497396518183, + "tpot_p50_ms": 15.280834627210176, + "tpot_p95_ms": 18.225302902624872, + "tpot_p99_ms": 18.700063700290844, + "e2e_p50_ms": 8596.120701506152, + "e2e_p95_ms": 10176.577183793415, + "e2e_p99_ms": 10183.192389139585, + "itl_p50_ms": 14.963374997023493, + "itl_p95_ms": 17.358059901744127, + "itl_p99_ms": 18.431273359310598, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i16384_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i16384_o1024_a1.json new file mode 100644 index 0000000..c24929f --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i16384_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 165.0628843260056, + "request_tps": 0.969327542368603, + "input_tps": 15881.462454167191, + "output_tps": 992.5914033854494, + "total_tps": 16874.05385755264, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 4267.802063499403, + "ttft_p95_ms": 19862.928376002674, + "ttft_p99_ms": 21657.044113324228, + "tpot_p50_ms": 25.376833950635163, + "tpot_p95_ms": 37.8741763598776, + "tpot_p99_ms": 40.72076095410008, + "e2e_p50_ms": 35063.07619450672, + "e2e_p95_ms": 51894.37677649694, + "e2e_p99_ms": 53688.66714284828, + "itl_p50_ms": 17.057024502719287, + "itl_p95_ms": 18.18839350307826, + "itl_p99_ms": 674.1142390163444, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i16384_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i16384_o2048_a1.json new file mode 100644 index 0000000..b985e62 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i16384_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 203.0028582619998, + "request_tps": 0.7881662424353681, + "input_tps": 12913.315716061072, + "output_tps": 1614.164464507634, + "total_tps": 14527.480180568706, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 2550.8961580053438, + "ttft_p95_ms": 15365.738819241233, + "ttft_p99_ms": 18318.440213190624, + "tpot_p50_ms": 16.89672226160163, + "tpot_p95_ms": 21.461009534389174, + "tpot_p99_ms": 22.790276562500463, + "e2e_p50_ms": 40052.36150848941, + "e2e_p95_ms": 48792.51678490764, + "e2e_p99_ms": 48817.223941341945, + "itl_p50_ms": 14.973600002122112, + "itl_p95_ms": 17.530049994093133, + "itl_p99_ms": 18.100957148271835, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i4096_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i4096_o1024_a1.json new file mode 100644 index 0000000..e3ac3a2 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i4096_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 86.79001112899277, + "request_tps": 1.843530124246648, + "input_tps": 7551.09938891427, + "output_tps": 1887.7748472285675, + "total_tps": 9438.874236142838, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 1054.5200394990388, + "ttft_p95_ms": 1617.5321672504651, + "ttft_p99_ms": 1622.538609701296, + "tpot_p50_ms": 15.406173739000165, + "tpot_p95_ms": 17.669181664998092, + "tpot_p99_ms": 18.132834118727168, + "e2e_p50_ms": 16632.105807497283, + "e2e_p95_ms": 18881.63431045614, + "e2e_p99_ms": 18892.737503780663, + "itl_p50_ms": 15.119137002329808, + "itl_p95_ms": 17.811838999477914, + "itl_p99_ms": 18.673117576690856, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i4096_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i4096_o2048_a1.json new file mode 100644 index 0000000..20ce709 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i4096_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 171.35218937399623, + "request_tps": 0.9337493765590661, + "input_tps": 3824.6374463859347, + "output_tps": 1912.3187231929674, + "total_tps": 5736.956169578902, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 897.8936719940975, + "ttft_p95_ms": 1402.46574630728, + "ttft_p99_ms": 1414.9006537273815, + "tpot_p50_ms": 17.125081185391775, + "tpot_p95_ms": 17.535651223030666, + "tpot_p99_ms": 17.64257730764133, + "e2e_p50_ms": 35692.707454494666, + "e2e_p95_ms": 36474.77745984506, + "e2e_p99_ms": 36482.97110349682, + "itl_p50_ms": 16.84335399477277, + "itl_p95_ms": 17.664605999016203, + "itl_p99_ms": 18.40053380583413, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i4096_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i4096_o256_a1.json new file mode 100644 index 0000000..424ea56 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i4096_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 39.068993989989394, + "request_tps": 4.0953191689808195, + "input_tps": 16774.427316145437, + "output_tps": 1048.4017072590898, + "total_tps": 17822.829023404527, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 2343.7372504995437, + "ttft_p95_ms": 5162.197779646522, + "ttft_p99_ms": 5491.549329505941, + "tpot_p50_ms": 19.213523033335253, + "tpot_p95_ms": 32.66720932803069, + "tpot_p99_ms": 34.66762047272469, + "e2e_p50_ms": 8660.10677999293, + "e2e_p95_ms": 9275.378481158987, + "e2e_p99_ms": 9288.702330155647, + "itl_p50_ms": 14.975201003835537, + "itl_p95_ms": 17.6850339921657, + "itl_p99_ms": 28.16523999499614, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i4096_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i4096_o4096_a1.json new file mode 100644 index 0000000..7c8679a --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i4096_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 361.17721821699524, + "request_tps": 0.442995825677665, + "input_tps": 1814.5109019757158, + "output_tps": 1814.5109019757158, + "total_tps": 3629.0218039514316, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 1892.501820497273, + "ttft_p95_ms": 4395.493246797548, + "ttft_p99_ms": 5224.942043306945, + "tpot_p50_ms": 17.487011898045463, + "tpot_p95_ms": 18.118140513893067, + "tpot_p99_ms": 18.395078147048498, + "e2e_p50_ms": 74226.02639300021, + "e2e_p95_ms": 75991.18075425722, + "e2e_p99_ms": 76000.55798450441, + "itl_p50_ms": 17.170556006021798, + "itl_p95_ms": 18.040875489532482, + "itl_p99_ms": 18.916477599123027, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i4096_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i4096_o512_a1.json new file mode 100644 index 0000000..e4c3a94 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c32_i4096_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 56.462142722986755, + "request_tps": 2.83375713856607, + "input_tps": 11607.069239566623, + "output_tps": 1450.8836549458279, + "total_tps": 13057.952894512451, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 1453.6274200072512, + "ttft_p95_ms": 3663.0988704906445, + "ttft_p99_ms": 4404.091445246595, + "tpot_p50_ms": 18.000749555776817, + "tpot_p95_ms": 23.179044860871425, + "tpot_p99_ms": 23.890643953811313, + "e2e_p50_ms": 11132.362890006334, + "e2e_p95_ms": 13523.678070852475, + "e2e_p99_ms": 13540.85523293179, + "itl_p50_ms": 17.186335011501797, + "itl_p95_ms": 18.29535180295352, + "itl_p99_ms": 19.591979318647628, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c48_i1024_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c48_i1024_o1024_a1.json new file mode 100644 index 0000000..5ff9e9f --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c48_i1024_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 97.32574620700325, + "request_tps": 2.46594564494313, + "input_tps": 2525.128340421765, + "output_tps": 2525.128340421765, + "total_tps": 5050.25668084353, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 971.1724649969256, + "ttft_p95_ms": 1047.563062050176, + "ttft_p99_ms": 1059.1346113267355, + "tpot_p50_ms": 17.463300068922933, + "tpot_p95_ms": 20.0403643846501, + "tpot_p99_ms": 20.23941422052442, + "e2e_p50_ms": 18865.876165502414, + "e2e_p95_ms": 20948.537042703538, + "e2e_p99_ms": 20966.204005898326, + "itl_p50_ms": 17.521906993351877, + "itl_p95_ms": 19.884526249370538, + "itl_p99_ms": 20.897341451927787, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c48_i1024_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c48_i1024_o2048_a1.json new file mode 100644 index 0000000..d3803fc --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c48_i1024_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 205.12268571999448, + "request_tps": 1.170031482171676, + "input_tps": 1198.1122377437962, + "output_tps": 2396.2244754875924, + "total_tps": 3594.3367132313883, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 1103.477151504194, + "ttft_p95_ms": 1655.3460868577529, + "ttft_p99_ms": 2063.1610584430746, + "tpot_p50_ms": 19.476409309720992, + "tpot_p95_ms": 19.98451202838762, + "tpot_p99_ms": 20.224875959688827, + "e2e_p50_ms": 41019.6206035107, + "e2e_p95_ms": 41983.93201486106, + "e2e_p99_ms": 42004.45509949772, + "itl_p50_ms": 19.33959749294445, + "itl_p95_ms": 21.063311200850876, + "itl_p99_ms": 22.709367697680136, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c48_i1024_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c48_i1024_o4096_a1.json new file mode 100644 index 0000000..5c887f3 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c48_i1024_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 387.6589750939893, + "request_tps": 0.6191008474441, + "input_tps": 633.9592677827584, + "output_tps": 2535.8370711310336, + "total_tps": 3169.7963389137917, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 1218.414176008082, + "ttft_p95_ms": 2534.173024399205, + "ttft_p99_ms": 2783.003787570633, + "tpot_p50_ms": 19.216241195482507, + "tpot_p95_ms": 19.885012049902812, + "tpot_p99_ms": 19.88576329422029, + "e2e_p50_ms": 79752.23233650468, + "e2e_p95_ms": 82442.38089720166, + "e2e_p99_ms": 82460.91719069882, + "itl_p50_ms": 18.868017999920994, + "itl_p95_ms": 20.22105274882051, + "itl_p99_ms": 21.84910236479482, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c48_i1024_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c48_i1024_o512_a1.json new file mode 100644 index 0000000..78a9d8b --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c48_i1024_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 57.3580158239929, + "request_tps": 4.184245158278432, + "input_tps": 4284.667042077114, + "output_tps": 2142.333521038557, + "total_tps": 6427.000563115672, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 1114.938983497268, + "ttft_p95_ms": 2623.5857654435677, + "ttft_p99_ms": 2657.5968347305025, + "tpot_p50_ms": 19.735206197671033, + "tpot_p95_ms": 22.902151353534165, + "tpot_p99_ms": 23.980957457851613, + "e2e_p50_ms": 11155.085417994997, + "e2e_p95_ms": 12560.576499555464, + "e2e_p99_ms": 12733.288235523505, + "itl_p50_ms": 19.239879009546712, + "itl_p95_ms": 20.936376793542877, + "itl_p99_ms": 22.45631740195677, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c48_i4096_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c48_i4096_o1024_a1.json new file mode 100644 index 0000000..5ea33ae --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c48_i4096_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 107.01144343199849, + "request_tps": 2.242750796577288, + "input_tps": 9186.307262780572, + "output_tps": 2296.576815695143, + "total_tps": 11482.884078475714, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 1429.1901315009454, + "ttft_p95_ms": 4655.2081622539845, + "ttft_p99_ms": 5421.287226663553, + "tpot_p50_ms": 17.961494444764135, + "tpot_p95_ms": 23.42196408294526, + "tpot_p99_ms": 23.725152669697756, + "e2e_p50_ms": 19863.534916497883, + "e2e_p95_ms": 25332.92310884499, + "e2e_p99_ms": 25341.139651168924, + "itl_p50_ms": 17.542485002195463, + "itl_p95_ms": 19.82468349451665, + "itl_p99_ms": 22.084278145339344, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c48_i4096_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c48_i4096_o2048_a1.json new file mode 100644 index 0000000..b810c05 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c48_i4096_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 195.2795769630029, + "request_tps": 1.2290071687602524, + "input_tps": 5034.013363241994, + "output_tps": 2517.006681620997, + "total_tps": 7551.02004486299, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 1402.5059129999136, + "ttft_p95_ms": 3014.964500940548, + "ttft_p99_ms": 3645.231739671726, + "tpot_p50_ms": 17.624665780901058, + "tpot_p95_ms": 20.07420706065267, + "tpot_p99_ms": 20.200438402067306, + "e2e_p50_ms": 38203.180718497606, + "e2e_p95_ms": 42234.80308325525, + "e2e_p99_ms": 42259.54538603983, + "itl_p50_ms": 17.31949599343352, + "itl_p95_ms": 20.04093239957001, + "itl_p99_ms": 21.313599089044146, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c48_i4096_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c48_i4096_o512_a1.json new file mode 100644 index 0000000..6e6a387 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c48_i4096_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 78.62980660999892, + "request_tps": 3.052277632964196, + "input_tps": 12502.129184621346, + "output_tps": 1562.7661480776683, + "total_tps": 14064.895332699014, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 2010.3960574997473, + "ttft_p95_ms": 8091.9554406515035, + "ttft_p99_ms": 9461.760722859906, + "tpot_p50_ms": 23.111368014671182, + "tpot_p95_ms": 34.88897108324944, + "tpot_p99_ms": 38.233606207983335, + "e2e_p50_ms": 14650.032247503987, + "e2e_p95_ms": 20735.955423545238, + "e2e_p99_ms": 21034.978971217053, + "itl_p50_ms": 19.645875494461507, + "itl_p95_ms": 22.24900509609142, + "itl_p99_ms": 25.97859771209185, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c64_i1024_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c64_i1024_o1024_a1.json new file mode 100644 index 0000000..d7abd07 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c64_i1024_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 320, + "failed": 0, + "duration_s": 111.44104069900641, + "request_tps": 2.8714735432549947, + "input_tps": 2940.3889082931146, + "output_tps": 2940.3889082931146, + "total_tps": 5880.777816586229, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 1251.2050785007887, + "ttft_p95_ms": 1664.9364830955165, + "ttft_p99_ms": 1687.642185555742, + "tpot_p50_ms": 20.892368739980142, + "tpot_p95_ms": 22.107194719502488, + "tpot_p99_ms": 22.406787545171895, + "e2e_p50_ms": 22625.58904299658, + "e2e_p95_ms": 23358.138840601896, + "e2e_p99_ms": 23385.716578069405, + "itl_p50_ms": 20.738980994792655, + "itl_p95_ms": 21.87385440629441, + "itl_p99_ms": 24.061159000266343, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c64_i1024_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c64_i1024_o2048_a1.json new file mode 100644 index 0000000..a94e6b3 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c64_i1024_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 320, + "failed": 0, + "duration_s": 227.52200377998815, + "request_tps": 1.4064573741598956, + "input_tps": 1440.212351139733, + "output_tps": 2880.424702279466, + "total_tps": 4320.637053419199, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 1391.4943349955138, + "ttft_p95_ms": 2766.4422078472853, + "ttft_p99_ms": 3516.292367083952, + "tpot_p50_ms": 21.495228763068177, + "tpot_p95_ms": 22.537871907642682, + "tpot_p99_ms": 22.736073239234297, + "e2e_p50_ms": 45078.62572550221, + "e2e_p95_ms": 47292.98801284676, + "e2e_p99_ms": 47307.55221481537, + "itl_p50_ms": 21.152492990950122, + "itl_p95_ms": 22.822722894488834, + "itl_p99_ms": 24.930158918141395, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c64_i1024_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c64_i1024_o4096_a1.json new file mode 100644 index 0000000..68692fe --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c64_i1024_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 320, + "failed": 0, + "duration_s": 429.61134696399677, + "request_tps": 0.7448592833066333, + "input_tps": 762.7359061059925, + "output_tps": 3050.94362442397, + "total_tps": 3813.6795305299624, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 1419.7698949938058, + "ttft_p95_ms": 2921.583319955971, + "ttft_p99_ms": 3675.6699006249255, + "tpot_p50_ms": 20.945350810012243, + "tpot_p95_ms": 21.617453843995357, + "tpot_p99_ms": 21.769692606291244, + "e2e_p50_ms": 87146.0334000003, + "e2e_p95_ms": 89824.40359709944, + "e2e_p99_ms": 89989.40303418975, + "itl_p50_ms": 20.697043000836857, + "itl_p95_ms": 22.144093887618506, + "itl_p99_ms": 24.082228924962692, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c64_i1024_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c64_i1024_o512_a1.json new file mode 100644 index 0000000..9664176 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c64_i1024_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 320, + "failed": 0, + "duration_s": 64.12715022399789, + "request_tps": 4.990086084945788, + "input_tps": 5109.848150984487, + "output_tps": 2554.9240754922434, + "total_tps": 7664.77222647673, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 1574.427007995837, + "ttft_p95_ms": 2990.9389156964608, + "ttft_p99_ms": 3418.0432158199255, + "tpot_p50_ms": 21.48998137866181, + "tpot_p95_ms": 26.06068395949076, + "tpot_p99_ms": 26.0992564425009, + "e2e_p50_ms": 13091.199074500764, + "e2e_p95_ms": 14394.905802008725, + "e2e_p99_ms": 14404.70782280463, + "itl_p50_ms": 20.648846999392845, + "itl_p95_ms": 22.798411999247037, + "itl_p99_ms": 25.05893900524825, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c64_i4096_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c64_i4096_o2048_a1.json new file mode 100644 index 0000000..cbef255 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c64_i4096_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 320, + "failed": 0, + "duration_s": 240.90383681698586, + "request_tps": 1.3283308569431518, + "input_tps": 5440.84319003915, + "output_tps": 2720.421595019575, + "total_tps": 8161.264785058724, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 1544.8923770018155, + "ttft_p95_ms": 9636.169896001238, + "ttft_p99_ms": 11990.489179552387, + "tpot_p50_ms": 21.333603897412154, + "tpot_p95_ms": 26.558363334025465, + "tpot_p99_ms": 27.72400532863237, + "e2e_p50_ms": 46171.67401249753, + "e2e_p95_ms": 58603.167127201596, + "e2e_p99_ms": 58643.71314777425, + "itl_p50_ms": 20.934874992235564, + "itl_p95_ms": 22.243883497139905, + "itl_p99_ms": 24.41439779868233, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c80_i1024_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c80_i1024_o1024_a1.json new file mode 100644 index 0000000..4bd7e38 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c80_i1024_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 400, + "failed": 0, + "duration_s": 123.47737352400145, + "request_tps": 3.2394598992846917, + "input_tps": 3317.2069368675243, + "output_tps": 3317.2069368675243, + "total_tps": 6634.413873735049, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 1259.8191590004717, + "ttft_p95_ms": 2176.3979704985104, + "ttft_p99_ms": 2561.1882165598217, + "tpot_p50_ms": 23.053262887589092, + "tpot_p95_ms": 23.774626320429117, + "tpot_p99_ms": 24.000006763245814, + "e2e_p50_ms": 24829.307881504064, + "e2e_p95_ms": 24994.919799604395, + "e2e_p99_ms": 25020.736574648326, + "itl_p50_ms": 22.634291992289945, + "itl_p95_ms": 24.973727005999535, + "itl_p99_ms": 27.328714399482124, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c80_i1024_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c80_i1024_o2048_a1.json new file mode 100644 index 0000000..ea05462 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c80_i1024_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 400, + "failed": 0, + "duration_s": 265.8902801889926, + "request_tps": 1.5043799258689836, + "input_tps": 1540.4850440898392, + "output_tps": 3080.9700881796784, + "total_tps": 4621.455132269518, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 1453.8656489967252, + "ttft_p95_ms": 4220.62352244684, + "ttft_p99_ms": 4392.5506810807565, + "tpot_p50_ms": 23.16868536492179, + "tpot_p95_ms": 32.49832417950609, + "tpot_p99_ms": 32.50333228816742, + "e2e_p50_ms": 49589.71686950099, + "e2e_p95_ms": 67762.65414134323, + "e2e_p99_ms": 67792.98369047901, + "itl_p50_ms": 22.87503899424337, + "itl_p95_ms": 24.68757720780559, + "itl_p99_ms": 27.605341025628153, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c80_i1024_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c80_i1024_o4096_a1.json new file mode 100644 index 0000000..4b22b8f --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c80_i1024_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 400, + "failed": 0, + "duration_s": 461.13784259100794, + "request_tps": 0.8674195935699159, + "input_tps": 888.2376638155939, + "output_tps": 3552.9506552623757, + "total_tps": 4441.18831907797, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 1903.3365680006682, + "ttft_p95_ms": 3811.413884991634, + "ttft_p99_ms": 4251.1653185045, + "tpot_p50_ms": 21.528175279001097, + "tpot_p95_ms": 23.315381058901853, + "tpot_p99_ms": 23.403718928096588, + "e2e_p50_ms": 90469.03869150265, + "e2e_p95_ms": 96639.56603370316, + "e2e_p99_ms": 96666.38061486068, + "itl_p50_ms": 21.778908499982208, + "itl_p95_ms": 24.338666697440196, + "itl_p99_ms": 26.772771790419934, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c80_i1024_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c80_i1024_o512_a1.json new file mode 100644 index 0000000..d6af508 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c80_i1024_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 400, + "failed": 0, + "duration_s": 69.2822109109984, + "request_tps": 5.773487807914353, + "input_tps": 5912.051515304297, + "output_tps": 2956.0257576521485, + "total_tps": 8868.077272956445, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 1404.177324497141, + "ttft_p95_ms": 3550.293400152441, + "ttft_p99_ms": 4256.765821308945, + "tpot_p50_ms": 23.07539692855886, + "tpot_p95_ms": 26.923700127783786, + "tpot_p99_ms": 28.125053756603382, + "e2e_p50_ms": 13326.003155001672, + "e2e_p95_ms": 14892.070211340615, + "e2e_p99_ms": 14897.991692002543, + "itl_p50_ms": 22.541024998645298, + "itl_p95_ms": 24.51956669683568, + "itl_p99_ms": 27.620439298043483, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i1024_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i1024_o256_a1.json new file mode 100644 index 0000000..6807d77 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i1024_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 18.495949040006963, + "request_tps": 2.1626357162576255, + "input_tps": 2214.5389734478085, + "output_tps": 553.6347433619521, + "total_tps": 2768.1737168097607, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 440.5348879954545, + "ttft_p95_ms": 3502.821924509044, + "ttft_p99_ms": 3505.5937448414625, + "tpot_p50_ms": 10.14652707840558, + "tpot_p95_ms": 22.57474098336602, + "tpot_p99_ms": 23.09809012723831, + "e2e_p50_ms": 3156.6136774999904, + "e2e_p95_ms": 6075.7029040003545, + "e2e_p99_ms": 6078.332349029661, + "itl_p50_ms": 10.086970491101965, + "itl_p95_ms": 10.408660549001068, + "itl_p99_ms": 11.378483117441656, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i131072_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i131072_o128_a1.json new file mode 100644 index 0000000..9ee9ba7 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i131072_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 380.95681663800497, + "request_tps": 0.1049987774283851, + "input_tps": 13762.399755093293, + "output_tps": 13.439843510833294, + "total_tps": 13775.839598604127, + "mean_input_tokens": 131072.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 40792.213654494844, + "ttft_p95_ms": 75418.06702326574, + "ttft_p99_ms": 76547.42356198141, + "tpot_p50_ms": 274.08619550392257, + "tpot_p95_ms": 467.0889975168508, + "tpot_p99_ms": 530.1239402170978, + "e2e_p50_ms": 77351.3717395108, + "e2e_p95_ms": 77859.17398448946, + "e2e_p99_ms": 82151.91505951836, + "itl_p50_ms": 10.688919486710802, + "itl_p95_ms": 1270.379202303593, + "itl_p99_ms": 1353.2490307680564, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i16384_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i16384_o128_a1.json new file mode 100644 index 0000000..a1c732d --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i16384_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 21.187404705997324, + "request_tps": 1.887914095900456, + "input_tps": 30931.58454723307, + "output_tps": 241.65300427525835, + "total_tps": 31173.237551508326, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 1964.1958890060778, + "ttft_p95_ms": 3285.6321635576023, + "ttft_p99_ms": 3311.663662179344, + "tpot_p50_ms": 17.89670796456699, + "tpot_p95_ms": 32.79861776727342, + "tpot_p99_ms": 35.39528190300389, + "e2e_p50_ms": 4186.461981502362, + "e2e_p95_ms": 4819.31074025124, + "e2e_p99_ms": 4820.3892862127395, + "itl_p50_ms": 12.01575299637625, + "itl_p95_ms": 12.53046725760214, + "itl_p99_ms": 338.20250410339213, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i16384_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i16384_o256_a1.json new file mode 100644 index 0000000..ed3c807 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i16384_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 25.08967021299759, + "request_tps": 1.5942816171126148, + "input_tps": 26120.71001477308, + "output_tps": 408.1360939808294, + "total_tps": 26528.84610875391, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 1583.4694790100912, + "ttft_p95_ms": 3495.4268617620983, + "ttft_p99_ms": 3574.4406940348563, + "tpot_p50_ms": 12.640562578470801, + "tpot_p95_ms": 19.0468859096262, + "tpot_p99_ms": 19.132231138046265, + "e2e_p50_ms": 4888.153373496607, + "e2e_p95_ms": 6651.8559770600405, + "e2e_p99_ms": 6667.568836012651, + "itl_p50_ms": 11.900605997652747, + "itl_p95_ms": 12.416587599727789, + "itl_p99_ms": 16.99327580136015, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i16384_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i16384_o512_a1.json new file mode 100644 index 0000000..7ea4114 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i16384_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 32.75996512599522, + "request_tps": 1.221002520795108, + "input_tps": 20004.90530070705, + "output_tps": 625.1532906470953, + "total_tps": 20630.058591354147, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 767.6595185039332, + "ttft_p95_ms": 1176.9666902473525, + "ttft_p99_ms": 1830.5755027475243, + "tpot_p50_ms": 11.52257645988111, + "tpot_p95_ms": 13.118042319375098, + "tpot_p99_ms": 13.18724245253204, + "e2e_p50_ms": 6451.215492495976, + "e2e_p95_ms": 7027.763503049937, + "e2e_p99_ms": 7034.832992453594, + "itl_p50_ms": 10.443512990605086, + "itl_p95_ms": 12.125153494707774, + "itl_p99_ms": 12.808226600463968, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i4096_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i4096_o128_a1.json new file mode 100644 index 0000000..9fa9e50 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i4096_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 13.312623847989016, + "request_tps": 3.004666882858133, + "input_tps": 12307.115552186913, + "output_tps": 384.59736100584104, + "total_tps": 12691.712913192754, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 910.605711498647, + "ttft_p95_ms": 1522.6708843554661, + "ttft_p99_ms": 1526.5052215589094, + "tpot_p50_ms": 12.926063618135059, + "tpot_p95_ms": 19.92139204688953, + "tpot_p99_ms": 20.694775890148065, + "e2e_p50_ms": 2659.4583239930216, + "e2e_p95_ms": 3033.2144388041343, + "e2e_p99_ms": 3039.971140041598, + "itl_p50_ms": 10.341610010073055, + "itl_p95_ms": 12.366353905235883, + "itl_p99_ms": 17.325819017132748, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i65536_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i65536_o1024_a1.json new file mode 100644 index 0000000..d9595a1 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i65536_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 251.80086339799163, + "request_tps": 0.15885569040633815, + "input_tps": 10410.766526469777, + "output_tps": 162.66822697609027, + "total_tps": 10573.434753445867, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 17590.520314995956, + "ttft_p95_ms": 38182.11513094574, + "ttft_p99_ms": 42436.39091802004, + "tpot_p50_ms": 30.034462025418435, + "tpot_p95_ms": 44.99948253655721, + "tpot_p99_ms": 45.840643626068875, + "e2e_p50_ms": 50083.116945505026, + "e2e_p95_ms": 56164.22145070171, + "e2e_p99_ms": 63398.186792482884, + "itl_p50_ms": 12.246457001310773, + "itl_p95_ms": 12.86962560552638, + "itl_p99_ms": 846.2615448146241, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i65536_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i65536_o128_a1.json new file mode 100644 index 0000000..746dc36 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i65536_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 198.8860901500011, + "request_tps": 0.20112014857264152, + "input_tps": 13180.610056856634, + "output_tps": 25.743379017298114, + "total_tps": 13206.353435873933, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 18747.548280000046, + "ttft_p95_ms": 37954.907318703044, + "ttft_p99_ms": 40759.29950434613, + "tpot_p50_ms": 156.92021938971945, + "tpot_p95_ms": 258.3804646850536, + "tpot_p99_ms": 286.1677625214699, + "e2e_p50_ms": 39171.36516800383, + "e2e_p95_ms": 45190.92419464505, + "e2e_p99_ms": 50389.63774079078, + "itl_p50_ms": 12.328808006714098, + "itl_p95_ms": 904.5139313006074, + "itl_p99_ms": 1091.3648403182742, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i65536_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i65536_o256_a1.json new file mode 100644 index 0000000..f15513a --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i65536_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 212.1054012759996, + "request_tps": 0.18858548513788426, + "input_tps": 12359.138353996383, + "output_tps": 48.27788419529837, + "total_tps": 12407.416238191681, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 17527.240303992585, + "ttft_p95_ms": 40766.78363170649, + "ttft_p99_ms": 45234.88084075011, + "tpot_p50_ms": 84.98415598624692, + "tpot_p95_ms": 150.79124716706013, + "tpot_p99_ms": 159.6644718746755, + "e2e_p50_ms": 40714.015546502196, + "e2e_p95_ms": 54715.512818601535, + "e2e_p99_ms": 68239.62552806508, + "itl_p50_ms": 12.285736011108384, + "itl_p95_ms": 823.8078383044922, + "itl_p99_ms": 1058.4284620790413, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i65536_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i65536_o512_a1.json new file mode 100644 index 0000000..831c1ae --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c8_i65536_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 221.21875818599074, + "request_tps": 0.1808164928146365, + "input_tps": 11849.989673100017, + "output_tps": 92.57804432109388, + "total_tps": 11942.56771742111, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 19337.339771998813, + "ttft_p95_ms": 38204.29200109792, + "ttft_p99_ms": 42388.65437057044, + "tpot_p50_ms": 48.08991940802787, + "tpot_p95_ms": 75.00450040077983, + "tpot_p99_ms": 78.97179943347363, + "e2e_p50_ms": 44083.91553600086, + "e2e_p95_ms": 48563.43825120639, + "e2e_p99_ms": 51796.121535643615, + "itl_p50_ms": 12.262899006600492, + "itl_p95_ms": 13.326331894495494, + "itl_p99_ms": 956.881430233477, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c96_i1024_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c96_i1024_o1024_a1.json new file mode 100644 index 0000000..a1f60ce --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c96_i1024_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 480, + "failed": 0, + "duration_s": 131.60086494100688, + "request_tps": 3.647392440886852, + "input_tps": 3734.9298594681363, + "output_tps": 3734.9298594681363, + "total_tps": 7469.859718936273, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 1641.8105130069307, + "ttft_p95_ms": 2725.6738467483956, + "ttft_p99_ms": 2848.544350732118, + "tpot_p50_ms": 24.315747637830274, + "tpot_p95_ms": 25.91790127140095, + "tpot_p99_ms": 26.00329389268108, + "e2e_p50_ms": 26516.128896502778, + "e2e_p95_ms": 27687.729706803657, + "e2e_p99_ms": 27709.92693845037, + "itl_p50_ms": 23.818820001906715, + "itl_p95_ms": 26.684747799299657, + "itl_p99_ms": 30.725880123209194, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c96_i1024_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c96_i1024_o4096_a1.json new file mode 100644 index 0000000..245e6c0 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c96_i1024_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 480, + "failed": 0, + "duration_s": 509.60229949399945, + "request_tps": 0.9419109773967023, + "input_tps": 964.5168408542231, + "output_tps": 3858.0673634168925, + "total_tps": 4822.584204271116, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 2537.5389005057514, + "ttft_p95_ms": 4154.587076739699, + "ttft_p99_ms": 4880.831743057352, + "tpot_p50_ms": 24.489344787912636, + "tpot_p95_ms": 25.11027214580131, + "tpot_p99_ms": 25.309028622957065, + "e2e_p50_ms": 102768.33627750602, + "e2e_p95_ms": 104733.21360939881, + "e2e_p99_ms": 104743.7613113386, + "itl_p50_ms": 24.15896699676523, + "itl_p95_ms": 25.35902190429624, + "itl_p99_ms": 28.550959882268216, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c96_i1024_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c96_i1024_o512_a1.json new file mode 100644 index 0000000..032e58e --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/metrics/vllm_c96_i1024_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 480, + "failed": 0, + "duration_s": 75.0480901490082, + "request_tps": 6.3958989368944446, + "input_tps": 6549.400511379911, + "output_tps": 3274.7002556899556, + "total_tps": 9824.100767069867, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 1664.6413179987576, + "ttft_p95_ms": 4356.379911445401, + "ttft_p99_ms": 4993.593463639991, + "tpot_p50_ms": 25.16758154697011, + "tpot_p95_ms": 30.006507990595942, + "tpot_p99_ms": 31.581847599314614, + "e2e_p50_ms": 14630.724068490963, + "e2e_p95_ms": 17554.996255355218, + "e2e_p99_ms": 17564.006839842623, + "itl_p50_ms": 23.962745995959267, + "itl_p95_ms": 26.025201004813425, + "itl_p99_ms": 29.676723000011375, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/server_cmd.txt b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/server_cmd.txt new file mode 100644 index 0000000..236c49e --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp4_dp2/server_cmd.txt @@ -0,0 +1 @@ +vllm serve /data1/hf_models/DeepSeek-V4-Flash --trust-remote-code --kv-cache-dtype fp8 --block-size 256 --tensor-parallel-size 4 --gpu-memory-utilization 0.9 --max-model-len 1048576 --max-num-seqs 128 --no-enable-flashinfer-autotune --host 0.0.0.0 --port 30030 --data-parallel-size 2 diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i1024_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i1024_o1024_a1.json new file mode 100644 index 0000000..72249ef --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i1024_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 61.41648891198565, + "request_tps": 1.3025817889825302, + "input_tps": 1333.843751918111, + "output_tps": 1333.843751918111, + "total_tps": 2667.687503836222, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 362.0326555101201, + "ttft_p95_ms": 606.7866882352973, + "ttft_p99_ms": 607.9597775882576, + "tpot_p50_ms": 11.630160520515538, + "tpot_p95_ms": 11.889058939988779, + "tpot_p99_ms": 11.943324029066188, + "e2e_p50_ms": 12311.110387017834, + "e2e_p95_ms": 12538.315513329871, + "e2e_p99_ms": 12539.58258541039, + "itl_p50_ms": 11.51326100807637, + "itl_p95_ms": 11.765323003055528, + "itl_p99_ms": 11.965371266705917, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i1024_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i1024_o128_a1.json new file mode 100644 index 0000000..e66748f --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i1024_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 52.02482991799479, + "request_tps": 1.5377272761122265, + "input_tps": 1574.63273073892, + "output_tps": 196.829091342365, + "total_tps": 1771.461822081285, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 1236.4054954814492, + "ttft_p95_ms": 35677.24587905686, + "ttft_p99_ms": 35680.42643753841, + "tpot_p50_ms": 15.189849878005925, + "tpot_p95_ms": 63.83978625250509, + "tpot_p99_ms": 164.92346879194884, + "e2e_p50_ms": 2717.587512001046, + "e2e_p95_ms": 37134.60250555654, + "e2e_p99_ms": 37136.07944580115, + "itl_p50_ms": 11.540331004653126, + "itl_p95_ms": 11.923325425595976, + "itl_p99_ms": 14.953987974149642, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i1024_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i1024_o2048_a1.json new file mode 100644 index 0000000..c0fe994 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i1024_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 123.58838067599572, + "request_tps": 0.6473100429216823, + "input_tps": 662.8454839518026, + "output_tps": 1325.6909679036053, + "total_tps": 1988.536451855408, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 503.70064900198486, + "ttft_p95_ms": 3332.6880413966137, + "ttft_p99_ms": 3334.5142474136082, + "tpot_p50_ms": 11.566837329992605, + "tpot_p95_ms": 11.80489444616316, + "tpot_p99_ms": 12.187286488214971, + "e2e_p50_ms": 24182.123156497255, + "e2e_p95_ms": 27096.394883874746, + "e2e_p99_ms": 27099.026713256026, + "itl_p50_ms": 11.55229500727728, + "itl_p95_ms": 11.839765001786873, + "itl_p99_ms": 12.08073129178956, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i1024_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i1024_o256_a1.json new file mode 100644 index 0000000..af35ed2 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i1024_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 61.86757407398545, + "request_tps": 1.2930844824193455, + "input_tps": 1324.1185099974098, + "output_tps": 331.02962749935244, + "total_tps": 1655.1481374967623, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 392.70260799094103, + "ttft_p95_ms": 27561.146233863838, + "ttft_p99_ms": 27563.441789576318, + "tpot_p50_ms": 12.039197992143112, + "tpot_p95_ms": 70.91977109907734, + "tpot_p99_ms": 90.06260683440567, + "e2e_p50_ms": 3515.55855650804, + "e2e_p95_ms": 30491.65073121112, + "e2e_p99_ms": 30493.552179136605, + "itl_p50_ms": 11.53830900148023, + "itl_p95_ms": 11.843036256323103, + "itl_p99_ms": 12.743973096075921, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i1024_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i1024_o4096_a1.json new file mode 100644 index 0000000..dff7d12 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i1024_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 238.9891159539984, + "request_tps": 0.3347432776620619, + "input_tps": 342.7771163259514, + "output_tps": 1371.1084653038056, + "total_tps": 1713.885581629757, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 407.33929100679234, + "ttft_p95_ms": 588.6433666979428, + "ttft_p99_ms": 591.8217892898247, + "tpot_p50_ms": 11.581026416729486, + "tpot_p95_ms": 11.690784101489584, + "tpot_p99_ms": 11.69744405043216, + "e2e_p50_ms": 47807.511286999215, + "e2e_p95_ms": 48180.03251795162, + "e2e_p99_ms": 48181.1768043015, + "itl_p50_ms": 11.548227979801595, + "itl_p95_ms": 11.81682851165533, + "itl_p99_ms": 11.96142429544125, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i1024_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i1024_o512_a1.json new file mode 100644 index 0000000..a9a2c0d --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i1024_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 32.14737898000749, + "request_tps": 2.488538802797956, + "input_tps": 2548.2637340651067, + "output_tps": 1274.1318670325534, + "total_tps": 3822.39560109766, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 413.3373150107218, + "ttft_p95_ms": 656.7111001349986, + "ttft_p99_ms": 658.238969738304, + "tpot_p50_ms": 11.758502294494425, + "tpot_p95_ms": 12.275672713020104, + "tpot_p99_ms": 12.393381951049646, + "e2e_p50_ms": 6405.572134506656, + "e2e_p95_ms": 6744.097330357181, + "e2e_p99_ms": 6745.546151806775, + "itl_p50_ms": 11.505689995829016, + "itl_p95_ms": 11.79091099766083, + "itl_p99_ms": 12.107418500818312, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i131072_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i131072_o128_a1.json new file mode 100644 index 0000000..7b490b7 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i131072_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 1061.9543217770115, + "request_tps": 0.07533280703272881, + "input_tps": 9874.02168339383, + "output_tps": 9.642599300189287, + "total_tps": 9883.664282694019, + "mean_input_tokens": 131072.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 109194.33135900181, + "ttft_p95_ms": 129901.68803883859, + "ttft_p99_ms": 173369.75827682696, + "tpot_p50_ms": 860.7348704526814, + "tpot_p95_ms": 863.007278436844, + "tpot_p99_ms": 877.5824045137604, + "e2e_p50_ms": 218323.68925350602, + "e2e_p95_ms": 239327.92834405843, + "e2e_p99_ms": 282901.98067860294, + "itl_p50_ms": 857.2086700005457, + "itl_p95_ms": 1084.6890412460198, + "itl_p99_ms": 1101.4063146195258, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i16384_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i16384_o1024_a1.json new file mode 100644 index 0000000..11cc5a3 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i16384_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 113.21908437900129, + "request_tps": 0.7065946561818122, + "input_tps": 11576.84684688281, + "output_tps": 723.5529279301757, + "total_tps": 12300.399774812988, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 1313.6420960072428, + "ttft_p95_ms": 15357.72944069031, + "ttft_p99_ms": 15506.10770056053, + "tpot_p50_ms": 15.984088203331268, + "tpot_p95_ms": 27.630180807318535, + "tpot_p99_ms": 30.014112934037232, + "e2e_p50_ms": 22529.161178987124, + "e2e_p95_ms": 37872.18460994045, + "e2e_p99_ms": 38033.13407269189, + "itl_p50_ms": 11.689694976666942, + "itl_p95_ms": 12.091477410285734, + "itl_p99_ms": 100.4964998154901, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i16384_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i16384_o128_a1.json new file mode 100644 index 0000000..9b82e9d --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i16384_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 86.4904059410037, + "request_tps": 0.924958082108773, + "input_tps": 15154.513217270136, + "output_tps": 118.39463450992294, + "total_tps": 15272.907851780057, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 7593.278043495957, + "ttft_p95_ms": 12824.893658250221, + "ttft_p99_ms": 14977.241600849084, + "tpot_p50_ms": 78.05374694479846, + "tpot_p95_ms": 140.04996209068491, + "tpot_p99_ms": 159.51224197268965, + "e2e_p50_ms": 17123.760913003935, + "e2e_p95_ms": 26093.928836856503, + "e2e_p99_ms": 26102.272943213466, + "itl_p50_ms": 11.707582991220988, + "itl_p95_ms": 648.0138543454814, + "itl_p99_ms": 666.6313856895431, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i16384_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i16384_o2048_a1.json new file mode 100644 index 0000000..575eb41 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i16384_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 125.30134890798945, + "request_tps": 0.6384608042707116, + "input_tps": 10460.541817171339, + "output_tps": 1307.5677271464174, + "total_tps": 11768.109544317756, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 779.641143992194, + "ttft_p95_ms": 1110.0852024552294, + "ttft_p99_ms": 1200.1118386295275, + "tpot_p50_ms": 11.868535543230175, + "tpot_p95_ms": 12.105919975430762, + "tpot_p99_ms": 12.135692873646816, + "e2e_p50_ms": 25057.22738750046, + "e2e_p95_ms": 25164.101334703446, + "e2e_p99_ms": 25175.635717739642, + "itl_p50_ms": 11.723723000613973, + "itl_p95_ms": 12.08025178639218, + "itl_p99_ms": 12.76948868646286, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i16384_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i16384_o256_a1.json new file mode 100644 index 0000000..d548f55 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i16384_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 20.272567973006517, + "request_tps": 3.946219349542801, + "input_tps": 64654.85782290925, + "output_tps": 1010.2321534829571, + "total_tps": 65665.0899763922, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 732.9621660028351, + "ttft_p95_ms": 1059.6370970582939, + "ttft_p99_ms": 1330.065922466165, + "tpot_p50_ms": 12.931638015616757, + "tpot_p95_ms": 14.641560330985836, + "tpot_p99_ms": 15.307604808072721, + "e2e_p50_ms": 3975.734623003518, + "e2e_p95_ms": 4291.209159711434, + "e2e_p99_ms": 4300.588087121723, + "itl_p50_ms": 11.725543983629905, + "itl_p95_ms": 12.27384575031465, + "itl_p99_ms": 21.68184940965146, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i16384_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i16384_o512_a1.json new file mode 100644 index 0000000..ff89253 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i16384_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 130.2362937349826, + "request_tps": 0.6142680945972843, + "input_tps": 10064.168461881905, + "output_tps": 314.50526443380954, + "total_tps": 10378.673726315716, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 8992.384835495614, + "ttft_p95_ms": 14412.819043596388, + "ttft_p99_ms": 18462.384867458015, + "tpot_p50_ms": 34.2774762993976, + "tpot_p95_ms": 47.27433437131593, + "tpot_p99_ms": 48.483385741755086, + "e2e_p50_ms": 26175.92193999735, + "e2e_p95_ms": 32332.358848227883, + "e2e_p99_ms": 36386.61295437544, + "itl_p50_ms": 11.672908993205056, + "itl_p95_ms": 13.615285683772521, + "itl_p99_ms": 648.6071743268986, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i4096_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i4096_o1024_a1.json new file mode 100644 index 0000000..f46b77b --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i4096_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 62.43282029798138, + "request_tps": 1.2813773207452974, + "input_tps": 5248.521505772738, + "output_tps": 1312.1303764431846, + "total_tps": 6560.651882215922, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 563.8070285058348, + "ttft_p95_ms": 777.6129835590837, + "ttft_p99_ms": 779.2977576740668, + "tpot_p50_ms": 11.56457677565038, + "tpot_p95_ms": 12.043608254741104, + "tpot_p99_ms": 12.209488064032353, + "e2e_p50_ms": 12494.566029999987, + "e2e_p95_ms": 12608.202403750329, + "e2e_p99_ms": 12608.88043828163, + "itl_p50_ms": 11.578381003346294, + "itl_p95_ms": 11.835356391384266, + "itl_p99_ms": 12.015851708129048, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i4096_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i4096_o128_a1.json new file mode 100644 index 0000000..786ad96 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i4096_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 31.652223498997046, + "request_tps": 2.527468567967585, + "input_tps": 10352.511254395227, + "output_tps": 323.51597669985085, + "total_tps": 10676.027231095079, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 2794.544570497237, + "ttft_p95_ms": 4019.1005029933863, + "ttft_p99_ms": 4515.16735956771, + "tpot_p50_ms": 28.070199511753664, + "tpot_p95_ms": 46.8207419802839, + "tpot_p99_ms": 47.12049638943226, + "e2e_p50_ms": 6348.330602006172, + "e2e_p95_ms": 6970.749582740245, + "e2e_p99_ms": 7275.766055090349, + "itl_p50_ms": 11.53999000962358, + "itl_p95_ms": 12.58592710510129, + "itl_p99_ms": 611.4497984829359, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i4096_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i4096_o2048_a1.json new file mode 100644 index 0000000..664c4e8 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i4096_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 121.8976368480071, + "request_tps": 0.6562883585655658, + "input_tps": 2688.1571166845574, + "output_tps": 1344.0785583422787, + "total_tps": 4032.235675026836, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 452.4103995063342, + "ttft_p95_ms": 869.4939421853633, + "ttft_p99_ms": 873.1171497041942, + "tpot_p50_ms": 11.663301339766598, + "tpot_p95_ms": 11.80714733549408, + "tpot_p99_ms": 11.878064581100528, + "e2e_p50_ms": 24396.55241399305, + "e2e_p95_ms": 24617.466119487653, + "e2e_p99_ms": 24621.38425450714, + "itl_p50_ms": 11.577293000300415, + "itl_p95_ms": 11.859655751322862, + "itl_p99_ms": 12.521442596334962, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i4096_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i4096_o256_a1.json new file mode 100644 index 0000000..4ceb101 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i4096_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 17.724754639988532, + "request_tps": 4.513461631762919, + "input_tps": 18487.138843700915, + "output_tps": 1155.4461777313072, + "total_tps": 19642.58502143222, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 584.0600009978516, + "ttft_p95_ms": 737.9011562516098, + "ttft_p99_ms": 741.6909766758907, + "tpot_p50_ms": 11.493453029349135, + "tpot_p95_ms": 13.185771285709194, + "tpot_p99_ms": 13.456647717705515, + "e2e_p50_ms": 3522.1422600006917, + "e2e_p95_ms": 3669.390212147846, + "e2e_p99_ms": 3672.551417663053, + "itl_p50_ms": 11.582269973587245, + "itl_p95_ms": 11.904317999142222, + "itl_p99_ms": 12.9585687990766, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i4096_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i4096_o4096_a1.json new file mode 100644 index 0000000..011aa2e --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i4096_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 239.23693925398402, + "request_tps": 0.3343965202425058, + "input_tps": 1369.6881469133039, + "output_tps": 1369.6881469133039, + "total_tps": 2739.3762938266077, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 535.376151005039, + "ttft_p95_ms": 764.6692137466744, + "ttft_p99_ms": 770.1709890385973, + "tpot_p50_ms": 11.550237434308961, + "tpot_p95_ms": 11.648775638909902, + "tpot_p99_ms": 11.68880080481187, + "e2e_p50_ms": 47866.3036479993, + "e2e_p95_ms": 48047.369737493864, + "e2e_p99_ms": 48052.40901362762, + "itl_p50_ms": 11.535027006175369, + "itl_p95_ms": 11.756750987842679, + "itl_p99_ms": 11.894877816084774, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i4096_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i4096_o512_a1.json new file mode 100644 index 0000000..4eebaf9 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i4096_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 32.73595824901713, + "request_tps": 2.4437958831524944, + "input_tps": 10009.787937392617, + "output_tps": 1251.2234921740771, + "total_tps": 11261.011429566694, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 580.2664404909592, + "ttft_p95_ms": 736.464851902565, + "ttft_p99_ms": 739.5049455872504, + "tpot_p50_ms": 11.585709298447135, + "tpot_p95_ms": 12.479319499847486, + "tpot_p99_ms": 12.621792366471547, + "e2e_p50_ms": 6533.585163517273, + "e2e_p95_ms": 6650.0653961644275, + "e2e_p99_ms": 6653.992265443085, + "itl_p50_ms": 11.581516009755433, + "itl_p95_ms": 11.963144004403148, + "itl_p99_ms": 12.631978923745926, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i65536_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i65536_o1024_a1.json new file mode 100644 index 0000000..d8852ab --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i65536_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 500.2035044890072, + "request_tps": 0.15993490505774363, + "input_tps": 10481.493937864287, + "output_tps": 163.77334277912948, + "total_tps": 10645.267280643417, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 30243.93685850373, + "ttft_p95_ms": 42969.91652948925, + "ttft_p99_ms": 61614.389155033525, + "tpot_p50_ms": 73.36121226490721, + "tpot_p95_ms": 91.33674583832028, + "tpot_p99_ms": 96.64777304192485, + "e2e_p50_ms": 105178.19468899688, + "e2e_p95_ms": 113684.78196981303, + "e2e_p99_ms": 132258.0071013729, + "itl_p50_ms": 12.48324201151263, + "itl_p95_ms": 713.5221524600638, + "itl_p99_ms": 825.3858603240224, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i65536_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i65536_o128_a1.json new file mode 100644 index 0000000..c9d2c5d --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i65536_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 467.49062519401195, + "request_tps": 0.17112642626106017, + "input_tps": 11214.94147144484, + "output_tps": 21.904182561415702, + "total_tps": 11236.845654006254, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 28092.216243006987, + "ttft_p95_ms": 65609.9809892679, + "ttft_p99_ms": 84234.906140567, + "tpot_p50_ms": 529.0584835078563, + "tpot_p95_ms": 689.1646421382568, + "tpot_p99_ms": 694.2038618913791, + "e2e_p50_ms": 94517.5517155003, + "e2e_p95_ms": 132926.59808922032, + "e2e_p99_ms": 151589.5588841254, + "itl_p50_ms": 700.4347535112174, + "itl_p95_ms": 837.8521792372339, + "itl_p99_ms": 848.7232003509416, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i65536_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i65536_o256_a1.json new file mode 100644 index 0000000..a92bf47 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i65536_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 452.2059691059985, + "request_tps": 0.17691053516643815, + "input_tps": 11594.00883266769, + "output_tps": 45.289097002608166, + "total_tps": 11639.2979296703, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 27964.375940995524, + "ttft_p95_ms": 42944.88978464359, + "ttft_p99_ms": 61584.43124415524, + "tpot_p50_ms": 264.55735394904644, + "tpot_p95_ms": 328.878599924461, + "tpot_p99_ms": 350.21582279077904, + "e2e_p50_ms": 95521.30952950392, + "e2e_p95_ms": 106008.97152810794, + "e2e_p99_ms": 121054.057937187, + "itl_p50_ms": 12.545981488074176, + "itl_p95_ms": 818.9031008194432, + "itl_p99_ms": 841.0442970949224, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i65536_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i65536_o512_a1.json new file mode 100644 index 0000000..784ff38 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c16_i65536_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 468.2541085879784, + "request_tps": 0.1708474064247727, + "input_tps": 11196.655627453903, + "output_tps": 87.47387208948362, + "total_tps": 11284.129499543387, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 28707.305190488114, + "ttft_p95_ms": 42941.14238748442, + "ttft_p99_ms": 61583.86625094918, + "tpot_p50_ms": 138.2095831995791, + "tpot_p95_ms": 170.5406067458202, + "tpot_p99_ms": 181.07553997671428, + "e2e_p50_ms": 98761.97971450165, + "e2e_p95_ms": 112307.20164615194, + "e2e_p99_ms": 130892.13807505231, + "itl_p50_ms": 12.49102050496731, + "itl_p95_ms": 781.4320894540288, + "itl_p99_ms": 836.5383420526632, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i1024_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i1024_o128_a1.json new file mode 100644 index 0000000..9837e60 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i1024_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 4.607907787984004, + "request_tps": 1.0850911585163339, + "input_tps": 1111.1333463207259, + "output_tps": 138.89166829009073, + "total_tps": 1250.0250146108165, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 68.60028099617921, + "ttft_p95_ms": 93.32109419628976, + "ttft_p99_ms": 98.21370284073053, + "tpot_p50_ms": 6.665402291183118, + "tpot_p95_ms": 6.668919995271301, + "tpot_p99_ms": 6.669149680904619, + "e2e_p50_ms": 914.933813008247, + "e2e_p95_ms": 940.2739335957449, + "e2e_p99_ms": 945.1957123156171, + "itl_p50_ms": 6.641707994276658, + "itl_p95_ms": 6.847297985223121, + "itl_p99_ms": 7.190719614736732, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i1024_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i1024_o256_a1.json new file mode 100644 index 0000000..3b0090f --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i1024_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 8.878280922974227, + "request_tps": 0.5631720874095746, + "input_tps": 576.6882175074044, + "output_tps": 144.1720543768511, + "total_tps": 720.8602718842554, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 71.51998899644241, + "ttft_p95_ms": 91.57086079358123, + "ttft_p99_ms": 95.35931215505116, + "tpot_p50_ms": 6.661176898028227, + "tpot_p95_ms": 6.664835591321154, + "tpot_p99_ms": 6.665386227228879, + "e2e_p50_ms": 1771.2285799789242, + "e2e_p95_ms": 1790.1660829898901, + "e2e_p99_ms": 1793.9073757897131, + "itl_p50_ms": 6.639207014814019, + "itl_p95_ms": 6.857857707655057, + "itl_p99_ms": 7.324532314087264, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i131072_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i131072_o128_a1.json new file mode 100644 index 0000000..6fbbc43 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i131072_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 62.04160683800001, + "request_tps": 0.08059107838802039, + "input_tps": 10563.233826474609, + "output_tps": 10.31565803366661, + "total_tps": 10573.549484508274, + "mean_input_tokens": 131072.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 14315.946837014053, + "ttft_p95_ms": 14424.979343393352, + "ttft_p99_ms": 14444.436821474228, + "tpot_p50_ms": 6.151024771691117, + "tpot_p95_ms": 6.183124977891795, + "tpot_p99_ms": 6.187243395545484, + "e2e_p50_ms": 15091.257207997842, + "e2e_p95_ms": 15205.586195399519, + "e2e_p99_ms": 15225.502308679279, + "itl_p50_ms": 6.912430500960909, + "itl_p95_ms": 7.0415690206573345, + "itl_p99_ms": 7.134565238084179, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i16384_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i16384_o512_a1.json new file mode 100644 index 0000000..d51cf93 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i16384_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 18.15804166198359, + "request_tps": 0.2753600907562737, + "input_tps": 4511.4997269507885, + "output_tps": 140.98436646721214, + "total_tps": 4652.484093418, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 186.52946400106885, + "ttft_p95_ms": 232.34625218901783, + "ttft_p99_ms": 237.30703042820096, + "tpot_p50_ms": 6.7337785733917475, + "tpot_p95_ms": 6.740160819146246, + "tpot_p99_ms": 6.740686006462676, + "e2e_p50_ms": 3627.490315004252, + "e2e_p95_ms": 3669.8924877971876, + "e2e_p99_ms": 3676.717489558505, + "itl_p50_ms": 6.754819012712687, + "itl_p95_ms": 6.876269215717912, + "itl_p99_ms": 7.019366837339476, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i65536_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i65536_o1024_a1.json new file mode 100644 index 0000000..bf4b5c5 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i65536_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 60.06458672598819, + "request_tps": 0.08324372600463838, + "input_tps": 5455.460827439981, + "output_tps": 85.2415754287497, + "total_tps": 5540.70240286873, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 6200.891699991189, + "ttft_p95_ms": 6272.143243986648, + "ttft_p99_ms": 6285.5985119822435, + "tpot_p50_ms": 6.7929081173096115, + "tpot_p95_ms": 6.796755083673428, + "tpot_p99_ms": 6.797365722498417, + "e2e_p50_ms": 13150.848732999293, + "e2e_p95_ms": 13210.16350100399, + "e2e_p99_ms": 13220.45055940398, + "itl_p50_ms": 6.836396001745015, + "itl_p95_ms": 6.958074710564688, + "itl_p99_ms": 7.039473712211475, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i65536_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i65536_o128_a1.json new file mode 100644 index 0000000..687a49e --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i65536_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 29.374553603003733, + "request_tps": 0.170215352633945, + "input_tps": 11155.23335021822, + "output_tps": 21.78756513714496, + "total_tps": 11177.020915355364, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 6167.725449980935, + "ttft_p95_ms": 6247.644395998213, + "ttft_p99_ms": 6256.791441601235, + "tpot_p50_ms": 6.4021568109673685, + "tpot_p95_ms": 6.590872122924481, + "tpot_p99_ms": 6.591545397903419, + "e2e_p50_ms": 7004.3386799807195, + "e2e_p95_ms": 7057.308506994741, + "e2e_p99_ms": 7065.896558194654, + "itl_p50_ms": 6.835169973783195, + "itl_p95_ms": 6.938054598867893, + "itl_p99_ms": 7.164961716625845, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i65536_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i65536_o256_a1.json new file mode 100644 index 0000000..b35a085 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i65536_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 33.741494465008145, + "request_tps": 0.14818549324142372, + "input_tps": 9711.484485069945, + "output_tps": 37.93548626980447, + "total_tps": 9749.419971339748, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 6187.758629996097, + "ttft_p95_ms": 6245.184306602459, + "ttft_p99_ms": 6254.763505321462, + "tpot_p50_ms": 6.614399407921797, + "tpot_p95_ms": 6.6246079560597515, + "tpot_p99_ms": 6.626020995101105, + "e2e_p50_ms": 7877.484064985765, + "e2e_p95_ms": 7930.040321213892, + "e2e_p99_ms": 7939.256353051169, + "itl_p50_ms": 6.8363494938239455, + "itl_p95_ms": 6.942608255485538, + "itl_p99_ms": 7.0092535417643385, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i65536_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i65536_o512_a1.json new file mode 100644 index 0000000..e36eeca --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c1_i65536_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 42.45334394200472, + "request_tps": 0.11777635247839305, + "input_tps": 7718.591036023967, + "output_tps": 60.30149246893724, + "total_tps": 7778.892528492904, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 6204.497470986098, + "ttft_p95_ms": 6222.823049785802, + "ttft_p99_ms": 6224.350896343822, + "tpot_p50_ms": 6.734576565566621, + "tpot_p95_ms": 6.741488577286328, + "tpot_p99_ms": 6.742628805848284, + "e2e_p50_ms": 9645.783593994565, + "e2e_p95_ms": 9660.80251399544, + "e2e_p99_ms": 9663.666046793805, + "itl_p50_ms": 6.839831999968737, + "itl_p95_ms": 6.9495914547587745, + "itl_p99_ms": 7.086883967567698, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i1024_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i1024_o1024_a1.json new file mode 100644 index 0000000..04c1d1c --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i1024_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 80.85768987398478, + "request_tps": 1.9787851996434358, + "input_tps": 2026.2760444348783, + "output_tps": 2026.2760444348783, + "total_tps": 4052.5520888697565, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 601.8352310056798, + "ttft_p95_ms": 875.1060022099409, + "ttft_p99_ms": 879.6998479493777, + "tpot_p50_ms": 15.16937012902533, + "tpot_p95_ms": 15.621438565148944, + "tpot_p99_ms": 15.709627801430935, + "e2e_p50_ms": 16193.191637503332, + "e2e_p95_ms": 16394.8902862423, + "e2e_p99_ms": 16400.64757927088, + "itl_p50_ms": 15.000293991761282, + "itl_p95_ms": 15.395940997404976, + "itl_p99_ms": 15.808226090157401, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i1024_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i1024_o2048_a1.json new file mode 100644 index 0000000..121ff79 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i1024_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 158.5647476890008, + "request_tps": 1.0090515220559253, + "input_tps": 1033.2687585852675, + "output_tps": 2066.537517170535, + "total_tps": 3099.8062757558023, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 669.473130008555, + "ttft_p95_ms": 865.0462039600825, + "ttft_p99_ms": 869.1866847060737, + "tpot_p50_ms": 15.184576202977428, + "tpot_p95_ms": 15.339674701513589, + "tpot_p99_ms": 15.58220753159747, + "e2e_p50_ms": 31597.606307492242, + "e2e_p95_ms": 32205.554889343333, + "e2e_p99_ms": 32214.89342865796, + "itl_p50_ms": 15.059143508551642, + "itl_p95_ms": 15.587451988540124, + "itl_p99_ms": 16.146498053567488, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i1024_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i1024_o4096_a1.json new file mode 100644 index 0000000..c3b2076 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i1024_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 315.3864424380008, + "request_tps": 0.507314134251199, + "input_tps": 519.4896734732278, + "output_tps": 2077.958693892911, + "total_tps": 2597.448367366139, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 827.0409834949533, + "ttft_p95_ms": 2127.0444475507247, + "ttft_p99_ms": 2149.356553926773, + "tpot_p50_ms": 15.163013353477309, + "tpot_p95_ms": 15.352130105484576, + "tpot_p99_ms": 15.488373952998492, + "e2e_p50_ms": 63035.64160001406, + "e2e_p95_ms": 63718.42900119955, + "e2e_p99_ms": 63728.90193682426, + "itl_p50_ms": 15.009783499408513, + "itl_p95_ms": 15.42943195818225, + "itl_p99_ms": 15.688629885262342, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i1024_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i1024_o512_a1.json new file mode 100644 index 0000000..5ac794c --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i1024_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 51.012810627988074, + "request_tps": 3.1364670566145265, + "input_tps": 3211.742265973275, + "output_tps": 1605.8711329866376, + "total_tps": 4817.613398959913, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 1028.030455505359, + "ttft_p95_ms": 1825.9766799485078, + "ttft_p99_ms": 2148.851445445325, + "tpot_p50_ms": 17.623297716239186, + "tpot_p95_ms": 19.62757935753655, + "tpot_p99_ms": 20.25833365699475, + "e2e_p50_ms": 10080.087392503629, + "e2e_p95_ms": 11066.638421798416, + "e2e_p99_ms": 11499.76811571978, + "itl_p50_ms": 16.898524991120212, + "itl_p95_ms": 17.623670147440862, + "itl_p99_ms": 18.77276543091283, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i16384_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i16384_o2048_a1.json new file mode 100644 index 0000000..e76d854 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i16384_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 262.4147839480138, + "request_tps": 0.6097217450663798, + "input_tps": 9989.681071167566, + "output_tps": 1248.7101338959458, + "total_tps": 11238.391205063512, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 2030.8082264964469, + "ttft_p95_ms": 23697.570598886523, + "ttft_p99_ms": 24982.915203587618, + "tpot_p50_ms": 23.179559257443874, + "tpot_p95_ms": 31.551512422186658, + "tpot_p99_ms": 33.554478466176256, + "e2e_p50_ms": 52189.317357493564, + "e2e_p95_ms": 86702.39336090744, + "e2e_p99_ms": 87993.08046271733, + "itl_p50_ms": 15.3286209970247, + "itl_p95_ms": 15.942659782012925, + "itl_p99_ms": 610.8087138849077, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i16384_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i16384_o256_a1.json new file mode 100644 index 0000000..bcca135 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i16384_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 125.27427309198538, + "request_tps": 1.2771975925377472, + "input_tps": 20925.60535613845, + "output_tps": 326.9625836896633, + "total_tps": 21252.567939828114, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 1955.9702225087676, + "ttft_p95_ms": 25687.426767007855, + "ttft_p99_ms": 26337.214467667218, + "tpot_p50_ms": 71.67799913920189, + "tpot_p95_ms": 145.99830021673134, + "tpot_p99_ms": 162.3011558566975, + "e2e_p50_ms": 24834.69372948457, + "e2e_p95_ms": 61469.757113161904, + "e2e_p99_ms": 62129.46268044296, + "itl_p50_ms": 15.310183982364833, + "itl_p95_ms": 637.7874485973734, + "itl_p99_ms": 661.2853236787487, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i4096_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i4096_o1024_a1.json new file mode 100644 index 0000000..062b6ad --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i4096_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 103.49840845502331, + "request_tps": 1.5459174917605643, + "input_tps": 6332.078046251271, + "output_tps": 1583.0195115628178, + "total_tps": 7915.097557814089, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 912.6817460055463, + "ttft_p95_ms": 22953.754147146537, + "ttft_p99_ms": 22989.648280964466, + "tpot_p50_ms": 15.027860665688419, + "tpot_p95_ms": 15.874237878655665, + "tpot_p99_ms": 34.52500240562136, + "e2e_p50_ms": 16423.773564994917, + "e2e_p95_ms": 38356.92362101836, + "e2e_p99_ms": 38391.64368835016, + "itl_p50_ms": 15.018527017673478, + "itl_p95_ms": 15.399335196707398, + "itl_p99_ms": 15.800710814073682, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i4096_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i4096_o2048_a1.json new file mode 100644 index 0000000..be495c4 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i4096_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 158.23376187801477, + "request_tps": 1.0111622077426614, + "input_tps": 4141.720402913941, + "output_tps": 2070.8602014569706, + "total_tps": 6212.580604370912, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 756.7443619918777, + "ttft_p95_ms": 1068.711685812741, + "ttft_p99_ms": 1076.5162225518725, + "tpot_p50_ms": 15.088843866877845, + "tpot_p95_ms": 15.323755214356261, + "tpot_p99_ms": 15.423469172625605, + "e2e_p50_ms": 31670.834326505428, + "e2e_p95_ms": 31954.730625603406, + "e2e_p99_ms": 31962.703760440054, + "itl_p50_ms": 14.991424977779388, + "itl_p95_ms": 15.389906987547874, + "itl_p99_ms": 15.759109035134308, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i4096_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i4096_o256_a1.json new file mode 100644 index 0000000..9835d00 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i4096_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 46.13867022198974, + "request_tps": 3.467806922700252, + "input_tps": 14204.137155380233, + "output_tps": 887.7585722112645, + "total_tps": 15091.895727591496, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 1084.7839874913916, + "ttft_p95_ms": 6128.039478484425, + "ttft_p99_ms": 6135.151792074903, + "tpot_p50_ms": 28.80620697841925, + "tpot_p95_ms": 44.439678273264136, + "tpot_p99_ms": 49.92359239300004, + "e2e_p50_ms": 9273.995624491363, + "e2e_p95_ms": 15928.954499101383, + "e2e_p99_ms": 15931.015178921225, + "itl_p50_ms": 14.997396006947383, + "itl_p95_ms": 15.599585502059199, + "itl_p99_ms": 613.4823884931393, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i4096_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i4096_o4096_a1.json new file mode 100644 index 0000000..157fe16 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i4096_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 333.05284896001103, + "request_tps": 0.4804042376446114, + "input_tps": 1967.7357573923282, + "output_tps": 1967.7357573923282, + "total_tps": 3935.4715147846564, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 1095.2641390031204, + "ttft_p95_ms": 6705.06853362167, + "ttft_p99_ms": 7314.000860454107, + "tpot_p50_ms": 15.522370325397588, + "tpot_p95_ms": 16.71266772969576, + "tpot_p99_ms": 17.02914903613981, + "e2e_p50_ms": 66306.44073350413, + "e2e_p95_ms": 73349.6555153004, + "e2e_p99_ms": 73959.53097622347, + "itl_p50_ms": 14.9437389918603, + "itl_p95_ms": 15.368272416526452, + "itl_p99_ms": 15.751762358704575, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i4096_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i4096_o512_a1.json new file mode 100644 index 0000000..d8ccb0a --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c32_i4096_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 43.90981093500159, + "request_tps": 3.6438325875928577, + "input_tps": 14925.138278780345, + "output_tps": 1865.6422848475431, + "total_tps": 16790.78056362789, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 799.3582674971549, + "ttft_p95_ms": 1175.5077986585093, + "ttft_p99_ms": 1183.6164445849136, + "tpot_p50_ms": 15.426591234855762, + "tpot_p95_ms": 17.008792841357067, + "tpot_p99_ms": 17.19561321969741, + "e2e_p50_ms": 8660.748706999584, + "e2e_p95_ms": 9361.31468063686, + "e2e_p99_ms": 9369.353665226081, + "itl_p50_ms": 15.0350559852086, + "itl_p95_ms": 17.14442181983031, + "itl_p99_ms": 17.799102468416095, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c48_i1024_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c48_i1024_o1024_a1.json new file mode 100644 index 0000000..ec8fd07 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c48_i1024_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 94.47741836198838, + "request_tps": 2.5402895650730493, + "input_tps": 2601.2565146348024, + "output_tps": 2601.2565146348024, + "total_tps": 5202.513029269605, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 872.004470002139, + "ttft_p95_ms": 1068.056017033814, + "ttft_p99_ms": 1074.3570806449861, + "tpot_p50_ms": 17.694852847022954, + "tpot_p95_ms": 18.194023427918026, + "tpot_p99_ms": 18.255092446847165, + "e2e_p50_ms": 18907.909675486735, + "e2e_p95_ms": 19168.525980616687, + "e2e_p99_ms": 19178.07574301085, + "itl_p50_ms": 17.358476005028933, + "itl_p95_ms": 17.918874911265448, + "itl_p99_ms": 18.629112805938348, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c48_i1024_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c48_i1024_o2048_a1.json new file mode 100644 index 0000000..ff2248f --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c48_i1024_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 187.52933345900965, + "request_tps": 1.279799781576355, + "input_tps": 1310.5149763341876, + "output_tps": 2621.0299526683752, + "total_tps": 3931.5449290025626, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 996.749361496768, + "ttft_p95_ms": 2177.432288348791, + "ttft_p99_ms": 2548.245648367156, + "tpot_p50_ms": 17.70298705446327, + "tpot_p95_ms": 18.496204302150304, + "tpot_p99_ms": 18.711075803456517, + "e2e_p50_ms": 37275.27416749217, + "e2e_p95_ms": 39046.80715519644, + "e2e_p99_ms": 39530.66943249112, + "itl_p50_ms": 17.395492002833635, + "itl_p95_ms": 17.924149986356497, + "itl_p99_ms": 18.54355659452267, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c48_i1024_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c48_i1024_o4096_a1.json new file mode 100644 index 0000000..437cad0 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c48_i1024_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 384.594265947002, + "request_tps": 0.6240342648090146, + "input_tps": 639.0110871644309, + "output_tps": 2556.0443486577237, + "total_tps": 3195.0554358221543, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 988.36895050772, + "ttft_p95_ms": 2441.5721719269623, + "ttft_p99_ms": 5020.39980761183, + "tpot_p50_ms": 17.599965752136836, + "tpot_p95_ms": 22.463060006479317, + "tpot_p99_ms": 22.464100362125627, + "e2e_p50_ms": 73127.70336250833, + "e2e_p95_ms": 92961.33970778319, + "e2e_p99_ms": 93251.61461467971, + "itl_p50_ms": 17.415053996955976, + "itl_p95_ms": 17.90137597708963, + "itl_p99_ms": 18.575296591734517, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c48_i1024_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c48_i1024_o512_a1.json new file mode 100644 index 0000000..ef08044 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c48_i1024_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 85.11057490797248, + "request_tps": 2.819861107265517, + "input_tps": 2887.5377738398893, + "output_tps": 1443.7688869199446, + "total_tps": 4331.306660759834, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 1176.0342860070523, + "ttft_p95_ms": 32154.562560931667, + "ttft_p99_ms": 32162.17692289414, + "tpot_p50_ms": 18.486646824851704, + "tpot_p95_ms": 74.38846726118646, + "tpot_p99_ms": 74.39255435188375, + "e2e_p50_ms": 11688.053546997253, + "e2e_p95_ms": 41145.71591668209, + "e2e_p99_ms": 41149.52881608595, + "itl_p50_ms": 17.28113798890263, + "itl_p95_ms": 17.941987194353715, + "itl_p99_ms": 19.419972189934924, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c48_i4096_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c48_i4096_o2048_a1.json new file mode 100644 index 0000000..1ff2246 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c48_i4096_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 206.79596914799185, + "request_tps": 1.1605642072657902, + "input_tps": 4753.670992960677, + "output_tps": 2376.8354964803384, + "total_tps": 7130.506489441016, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 1145.1450809981907, + "ttft_p95_ms": 6598.897911023221, + "ttft_p99_ms": 9627.660050078412, + "tpot_p50_ms": 17.8303520544688, + "tpot_p95_ms": 22.353743953683303, + "tpot_p99_ms": 23.511489080978755, + "e2e_p50_ms": 37381.53092398716, + "e2e_p95_ms": 50990.53589172253, + "e2e_p99_ms": 54006.87992766645, + "itl_p50_ms": 17.39579500281252, + "itl_p95_ms": 17.878196993842717, + "itl_p99_ms": 18.62087064306253, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c48_i4096_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c48_i4096_o512_a1.json new file mode 100644 index 0000000..2de9da6 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c48_i4096_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 76.34642335801618, + "request_tps": 3.1435657290001995, + "input_tps": 12876.045225984817, + "output_tps": 1609.5056532481021, + "total_tps": 14485.55087923292, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 1696.7760205006925, + "ttft_p95_ms": 7079.009777032479, + "ttft_p99_ms": 9906.125367596096, + "tpot_p50_ms": 18.908712023488928, + "tpot_p95_ms": 37.75210648097284, + "tpot_p99_ms": 42.41224298877765, + "e2e_p50_ms": 13078.812665509759, + "e2e_p95_ms": 24145.498285339272, + "e2e_p99_ms": 26969.445044811862, + "itl_p50_ms": 17.41874399886001, + "itl_p95_ms": 17.961374601873104, + "itl_p99_ms": 383.83775273512606, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c64_i1024_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c64_i1024_o1024_a1.json new file mode 100644 index 0000000..01e65be --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c64_i1024_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 320, + "failed": 0, + "duration_s": 111.18827333900845, + "request_tps": 2.8780013430403155, + "input_tps": 2947.073375273283, + "output_tps": 2947.073375273283, + "total_tps": 5894.146750546566, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 1236.4594759856118, + "ttft_p95_ms": 3486.339133232832, + "ttft_p99_ms": 3954.5976989151677, + "tpot_p50_ms": 20.184367791299923, + "tpot_p95_ms": 22.733724067629, + "tpot_p99_ms": 23.26845901167211, + "e2e_p50_ms": 21600.943245503004, + "e2e_p95_ms": 24639.851032342995, + "e2e_p99_ms": 24654.7415991014, + "itl_p50_ms": 19.554279977455735, + "itl_p95_ms": 20.101924979826435, + "itl_p99_ms": 21.4101209945511, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c64_i1024_o2048_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c64_i1024_o2048_a1.json new file mode 100644 index 0000000..1680863 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c64_i1024_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 320, + "failed": 0, + "duration_s": 231.3560413280211, + "request_tps": 1.3831495307541928, + "input_tps": 1416.3451194922934, + "output_tps": 2832.690238984587, + "total_tps": 4249.03535847688, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 1237.1772454935126, + "ttft_p95_ms": 4075.03496848949, + "ttft_p99_ms": 4438.276356909774, + "tpot_p50_ms": 19.961317755979394, + "tpot_p95_ms": 29.797527954492573, + "tpot_p99_ms": 29.79867891234008, + "e2e_p50_ms": 41953.71764700394, + "e2e_p95_ms": 61770.79950050393, + "e2e_p99_ms": 61851.04672904883, + "itl_p50_ms": 19.631187984487042, + "itl_p95_ms": 20.16204949177336, + "itl_p99_ms": 20.984756899997592, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c64_i1024_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c64_i1024_o4096_a1.json new file mode 100644 index 0000000..b2e0883 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c64_i1024_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 320, + "failed": 0, + "duration_s": 413.36956618897966, + "request_tps": 0.7741256884250303, + "input_tps": 792.704704947231, + "output_tps": 3170.818819788924, + "total_tps": 3963.523524736155, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 1353.9671769976849, + "ttft_p95_ms": 3372.6759848868824, + "ttft_p99_ms": 3923.934282742266, + "tpot_p50_ms": 19.719561329914566, + "tpot_p95_ms": 20.46792597871924, + "tpot_p99_ms": 20.599813236414754, + "e2e_p50_ms": 82000.917450001, + "e2e_p95_ms": 85134.73726001248, + "e2e_p99_ms": 85152.08560849074, + "itl_p50_ms": 19.557946507120505, + "itl_p95_ms": 20.046211070439313, + "itl_p99_ms": 20.78921643726062, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c80_i1024_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c80_i1024_o1024_a1.json new file mode 100644 index 0000000..faa113b --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c80_i1024_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 400, + "failed": 0, + "duration_s": 131.46577305399114, + "request_tps": 3.0426170303332536, + "input_tps": 3115.6398390612517, + "output_tps": 3115.6398390612517, + "total_tps": 6231.279678122503, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 1146.1891874932917, + "ttft_p95_ms": 4487.726777615899, + "ttft_p99_ms": 5507.241578921676, + "tpot_p50_ms": 23.907261485326536, + "tpot_p95_ms": 26.730242256350078, + "tpot_p99_ms": 27.79943066497788, + "e2e_p50_ms": 25425.060303488863, + "e2e_p95_ms": 29093.009902854097, + "e2e_p99_ms": 29212.41997377714, + "itl_p50_ms": 23.131843001465313, + "itl_p95_ms": 27.57070749357808, + "itl_p99_ms": 28.485130463086534, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c80_i1024_o4096_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c80_i1024_o4096_a1.json new file mode 100644 index 0000000..0601b4d --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c80_i1024_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 400, + "failed": 0, + "duration_s": 501.7031515800045, + "request_tps": 0.797284208281904, + "input_tps": 816.4190292806697, + "output_tps": 3265.676117122679, + "total_tps": 4082.095146403349, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 2250.2203854965046, + "ttft_p95_ms": 4658.4656137551065, + "ttft_p99_ms": 5716.611436419189, + "tpot_p50_ms": 24.059723582905082, + "tpot_p95_ms": 24.354225822548287, + "tpot_p99_ms": 24.487317885165243, + "e2e_p50_ms": 99781.74352001224, + "e2e_p95_ms": 103402.10737639281, + "e2e_p99_ms": 104522.76062945835, + "itl_p50_ms": 23.14190100878477, + "itl_p95_ms": 23.698028920625802, + "itl_p99_ms": 24.58483098977012, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i1024_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i1024_o128_a1.json new file mode 100644 index 0000000..ce3ca72 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i1024_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 73.70493139498285, + "request_tps": 0.5427045279459121, + "input_tps": 555.729436616614, + "output_tps": 69.46617957707674, + "total_tps": 625.1956161936906, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 3509.9659979896387, + "ttft_p95_ms": 27374.682503790245, + "ttft_p99_ms": 32961.56131610186, + "tpot_p50_ms": 9.04847947639087, + "tpot_p95_ms": 249.60066149328418, + "tpot_p99_ms": 278.0204192110943, + "e2e_p50_ms": 4659.724333003396, + "e2e_p95_ms": 37692.61776938074, + "e2e_p99_ms": 37693.20461160998, + "itl_p50_ms": 9.10900448798202, + "itl_p95_ms": 9.500342964020094, + "itl_p99_ms": 10.415470011648722, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i1024_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i1024_o256_a1.json new file mode 100644 index 0000000..53ed074 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i1024_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 23.89860760301235, + "request_tps": 1.6737376781297548, + "input_tps": 1713.907382404869, + "output_tps": 428.47684560121724, + "total_tps": 2142.3842280060862, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 240.56512699462473, + "ttft_p95_ms": 10569.026969805418, + "ttft_p99_ms": 10732.348576473596, + "tpot_p50_ms": 9.789249927430944, + "tpot_p95_ms": 10.445169046887552, + "tpot_p99_ms": 35.187820401407905, + "e2e_p50_ms": 2737.936751509551, + "e2e_p95_ms": 13064.727992002736, + "e2e_p99_ms": 13066.108254613355, + "itl_p50_ms": 9.119076014030725, + "itl_p95_ms": 9.404584451112896, + "itl_p99_ms": 10.967337382608093, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i131072_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i131072_o128_a1.json new file mode 100644 index 0000000..5bfb77d --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i131072_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 542.2442093600112, + "request_tps": 0.07376750052750287, + "input_tps": 9668.853829140857, + "output_tps": 9.442240067520368, + "total_tps": 9678.296069208376, + "mean_input_tokens": 131072.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 56039.164249494206, + "ttft_p95_ms": 72600.3135224935, + "ttft_p99_ms": 92221.81805258355, + "tpot_p50_ms": 387.13930614164747, + "tpot_p95_ms": 753.8479004563203, + "tpot_p99_ms": 754.9241562044031, + "e2e_p50_ms": 110909.34608949465, + "e2e_p95_ms": 112789.03825180455, + "e2e_p99_ms": 134558.94819398702, + "itl_p50_ms": 660.2215159800835, + "itl_p95_ms": 1072.356289398158, + "itl_p99_ms": 1163.3652878389696, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i16384_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i16384_o1024_a1.json new file mode 100644 index 0000000..eae32b9 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i16384_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 50.61696275198483, + "request_tps": 0.7902489170674606, + "input_tps": 12947.438257233274, + "output_tps": 809.2148910770796, + "total_tps": 13756.653148310354, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 573.4946330048842, + "ttft_p95_ms": 733.3654505957384, + "ttft_p99_ms": 790.053364574851, + "tpot_p50_ms": 9.346392439878139, + "tpot_p95_ms": 9.664219554327293, + "tpot_p99_ms": 9.85550387294379, + "e2e_p50_ms": 10074.975329000154, + "e2e_p95_ms": 10331.053190038074, + "e2e_p99_ms": 10332.716352115385, + "itl_p50_ms": 9.268547495594248, + "itl_p95_ms": 9.488147572847083, + "itl_p99_ms": 10.121553320495892, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i16384_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i16384_o128_a1.json new file mode 100644 index 0000000..412718e --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i16384_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 8.840474782977253, + "request_tps": 4.524643866076273, + "input_tps": 74131.76510179366, + "output_tps": 579.154414857763, + "total_tps": 74710.91951665143, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 611.7119685077341, + "ttft_p95_ms": 705.7421804507612, + "ttft_p99_ms": 728.9413817677995, + "tpot_p50_ms": 8.631740236223134, + "tpot_p95_ms": 11.94178506870173, + "tpot_p99_ms": 12.782024573981596, + "e2e_p50_ms": 1712.7858160092728, + "e2e_p95_ms": 1891.0789757341263, + "e2e_p99_ms": 1892.5092905788915, + "itl_p50_ms": 9.147118500550278, + "itl_p95_ms": 9.401795730809681, + "itl_p99_ms": 14.493186495383268, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i16384_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i16384_o512_a1.json new file mode 100644 index 0000000..1615f5e --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i16384_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 73.81995394398109, + "request_tps": 0.5418589129756752, + "input_tps": 8877.816430193463, + "output_tps": 277.4317634435457, + "total_tps": 9155.248193637008, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 5848.043638005038, + "ttft_p95_ms": 10270.304282526195, + "ttft_p99_ms": 10305.013074664457, + "tpot_p50_ms": 17.2701541448243, + "tpot_p95_ms": 26.505919335452827, + "tpot_p99_ms": 26.557917271059214, + "e2e_p50_ms": 14963.280752504943, + "e2e_p95_ms": 15029.205415995966, + "e2e_p99_ms": 15033.096053168701, + "itl_p50_ms": 9.254616481484845, + "itl_p95_ms": 9.525524696800858, + "itl_p99_ms": 610.016989769938, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i4096_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i4096_o128_a1.json new file mode 100644 index 0000000..546dac9 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i4096_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 8.121882901003119, + "request_tps": 4.924966351713797, + "input_tps": 20172.662176619713, + "output_tps": 630.395693019366, + "total_tps": 20803.05786963908, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 391.37257999391295, + "ttft_p95_ms": 654.7860209611827, + "ttft_p99_ms": 655.5212389715598, + "tpot_p50_ms": 9.02552549213017, + "tpot_p95_ms": 11.834378843241568, + "tpot_p99_ms": 12.646475187780089, + "e2e_p50_ms": 1594.1091245040298, + "e2e_p95_ms": 1801.0780599128339, + "e2e_p99_ms": 1801.8622757561388, + "itl_p50_ms": 9.091041501960717, + "itl_p95_ms": 9.295591751288157, + "itl_p99_ms": 11.05267516104506, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i65536_o1024_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i65536_o1024_a1.json new file mode 100644 index 0000000..f9ce42b --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i65536_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 278.18836335701053, + "request_tps": 0.14378746658308766, + "input_tps": 9423.255409989233, + "output_tps": 147.23836578108177, + "total_tps": 9570.493775770316, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 24218.15920100198, + "ttft_p95_ms": 37847.32255409762, + "ttft_p99_ms": 40382.343416879594, + "tpot_p50_ms": 29.127336590921143, + "tpot_p95_ms": 49.198489598973694, + "tpot_p99_ms": 49.27219220356028, + "e2e_p50_ms": 56596.45105450181, + "e2e_p95_ms": 56909.13160444033, + "e2e_p99_ms": 62302.39748750086, + "itl_p50_ms": 9.62444901233539, + "itl_p95_ms": 9.987523194286041, + "itl_p99_ms": 776.7235341103515, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i65536_o128_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i65536_o128_a1.json new file mode 100644 index 0000000..d31da75 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i65536_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 235.13304557901574, + "request_tps": 0.17011645428867683, + "input_tps": 11148.751948262725, + "output_tps": 21.774906148950635, + "total_tps": 11170.526854411675, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 24172.63857051148, + "ttft_p95_ms": 35960.15721101285, + "ttft_p99_ms": 39592.695025814464, + "tpot_p50_ms": 166.83349701175814, + "tpot_p95_ms": 328.55576824153474, + "tpot_p99_ms": 329.2510806872486, + "e2e_p50_ms": 48098.485034002806, + "e2e_p95_ms": 51191.40079334465, + "e2e_p99_ms": 56645.06894614664, + "itl_p50_ms": 9.62428402272053, + "itl_p95_ms": 816.3079831079814, + "itl_p99_ms": 889.5722046121956, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i65536_o256_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i65536_o256_a1.json new file mode 100644 index 0000000..a57663f --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i65536_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 241.68815920100315, + "request_tps": 0.1655025224745639, + "input_tps": 10846.37331289302, + "output_tps": 42.36864575348836, + "total_tps": 10888.741958646508, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 24228.64825501165, + "ttft_p95_ms": 35330.018459535495, + "ttft_p99_ms": 39612.372868892846, + "tpot_p50_ms": 88.08975900584996, + "tpot_p95_ms": 168.72122898116345, + "tpot_p99_ms": 169.04595058763002, + "e2e_p50_ms": 49360.89754800196, + "e2e_p95_ms": 52881.204059695294, + "e2e_p99_ms": 57884.091986017, + "itl_p50_ms": 9.613202011678368, + "itl_p95_ms": 762.7214051928601, + "itl_p99_ms": 843.184829771053, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i65536_o512_a1.json b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i65536_o512_a1.json new file mode 100644 index 0000000..b06ce9a --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/metrics/vllm_c8_i65536_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 253.59146945900284, + "request_tps": 0.15773401244660815, + "input_tps": 10337.256239700911, + "output_tps": 80.75981437266337, + "total_tps": 10418.016054073574, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 18460.19749098923, + "ttft_p95_ms": 35239.79678586038, + "ttft_p99_ms": 39810.581101148564, + "tpot_p50_ms": 60.03944050394105, + "tpot_p95_ms": 88.83629755211601, + "tpot_p99_ms": 89.054011180707, + "e2e_p50_ms": 51667.23371298576, + "e2e_p95_ms": 64968.29380615088, + "e2e_p99_ms": 74357.24756326554, + "itl_p50_ms": 9.60926100378856, + "itl_p95_ms": 689.5956993947038, + "itl_p99_ms": 821.5999770630151, + "validation_errors": [] +} diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/server_cmd.txt b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/server_cmd.txt new file mode 100644 index 0000000..3450120 --- /dev/null +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/adaptive_results/adaptive_20260716-051910/tp8_dp1/server_cmd.txt @@ -0,0 +1 @@ +vllm serve /data1/hf_models/DeepSeek-V4-Flash --trust-remote-code --kv-cache-dtype fp8 --block-size 256 --tensor-parallel-size 8 --gpu-memory-utilization 0.9 --max-model-len 1048576 --max-num-seqs 128 --no-enable-flashinfer-autotune --host 0.0.0.0 --port 30030 diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/run_adaptive_concurrency.sh b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/run_adaptive_concurrency.sh index 3454a44..f76a93d 100755 --- a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/run_adaptive_concurrency.sh +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/run_adaptive_concurrency.sh @@ -6,15 +6,15 @@ SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" EXPERIMENT_NAME="$(basename "$SCRIPT_DIR")" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/lib.sh" +source "${SCRIPT_DIR}/../../../scripts/common/lib.sh" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/platform.sh" +source "${SCRIPT_DIR}/../../../scripts/common/platform.sh" # shellcheck source=/dev/null source "${SCRIPT_DIR}/config.env" # shellcheck source=/dev/null source "${SCRIPT_DIR}/adaptive_config.env" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/adaptive_bench_lib.sh" +source "${SCRIPT_DIR}/../../../scripts/common/adaptive_bench_lib.sh" ENGINE="vllm" ENGINE_PORT="$VLLM_PORT" diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/run_adaptive_concurrency_add16.sh b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/run_adaptive_concurrency_add16.sh index 7c905e8..fb18812 100755 --- a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/run_adaptive_concurrency_add16.sh +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/run_adaptive_concurrency_add16.sh @@ -6,15 +6,15 @@ SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" EXPERIMENT_NAME="$(basename "$SCRIPT_DIR")" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/lib.sh" +source "${SCRIPT_DIR}/../../../scripts/common/lib.sh" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/platform.sh" +source "${SCRIPT_DIR}/../../../scripts/common/platform.sh" # shellcheck source=/dev/null source "${SCRIPT_DIR}/config.env" # shellcheck source=/dev/null source "${SCRIPT_DIR}/adaptive_config.env" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/adaptive_bench_lib.sh" +source "${SCRIPT_DIR}/../../../scripts/common/adaptive_bench_lib.sh" ENGINE="vllm" ENGINE_PORT="$VLLM_PORT" diff --git a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/run_bench.sh b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/run_bench.sh index 3523651..1db82ea 100755 --- a/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/run_bench.sh +++ b/experiments/h20/dsv4_h20_vllm_tp_dp_matrix/run_bench.sh @@ -6,9 +6,9 @@ SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" EXPERIMENT_NAME="$(basename "$SCRIPT_DIR")" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/lib.sh" +source "${SCRIPT_DIR}/../../../scripts/common/lib.sh" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/platform.sh" +source "${SCRIPT_DIR}/../../../scripts/common/platform.sh" # shellcheck source=/dev/null source "${SCRIPT_DIR}/config.env" @@ -512,7 +512,7 @@ run_parallel_config() { # Parse results. log "parsing ${config_label} results" - "$PYTHON" "${SCRIPT_DIR}/../../../../scripts/common/parse_backend.py" "$result_root" --backend vllm \ + "$PYTHON" "${SCRIPT_DIR}/../../../scripts/common/parse_backend.py" "$result_root" --backend vllm \ >> "${phase_log_dir}/parse.log" 2>&1 || { log "WARNING: parser failed for ${config_label}; see ${phase_log_dir}/parse.log" } diff --git a/experiments/p800/dsv4_p800_256k_4k_probe/run_bench.sh b/experiments/p800/dsv4_p800_256k_4k_probe/run_bench.sh index 27ee8f4..adfe559 100755 --- a/experiments/p800/dsv4_p800_256k_4k_probe/run_bench.sh +++ b/experiments/p800/dsv4_p800_256k_4k_probe/run_bench.sh @@ -5,13 +5,13 @@ SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" EXPERIMENT_NAME="$(basename "$SCRIPT_DIR")" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/lib.sh" +source "${SCRIPT_DIR}/../../../scripts/common/lib.sh" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/platform.sh" +source "${SCRIPT_DIR}/../../../scripts/common/platform.sh" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/server_docker.sh" +source "${SCRIPT_DIR}/../../../scripts/common/server_docker.sh" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/bench_client_docker.sh" +source "${SCRIPT_DIR}/../../../scripts/common/bench_client_docker.sh" # shellcheck source=/dev/null source "${SCRIPT_DIR}/config.env" diff --git a/experiments/p800/dsv4_p800_256k_4k_probe/start_server.sh b/experiments/p800/dsv4_p800_256k_4k_probe/start_server.sh index 0fbdbaf..9e58b8d 100755 --- a/experiments/p800/dsv4_p800_256k_4k_probe/start_server.sh +++ b/experiments/p800/dsv4_p800_256k_4k_probe/start_server.sh @@ -5,7 +5,7 @@ SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" EXPERIMENT_NAME="$(basename "$SCRIPT_DIR")" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/server_docker.sh" +source "${SCRIPT_DIR}/../../../scripts/common/server_docker.sh" # shellcheck source=/dev/null source "${SCRIPT_DIR}/config.env" diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp2_dp4/server_cmd.txt b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp2_dp4/server_cmd.txt new file mode 100644 index 0000000..cddc8dc --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp2_dp4/server_cmd.txt @@ -0,0 +1 @@ +--host 0.0.0.0 --port 30014 --model-path /models --attention-backend nsa --nsa-prefill klxdsa --nsa-decode klxdsa --trust-remote-code --disable-custom-all-reduce --page-size 64 --mem-fraction-static 0.8 --tensor-parallel-size 2 --disable-shared-experts-fusion --quantization w8a8_int8 --kv-cache-dtype float16 --disable-piecewise-cuda-graph --cuda-graph-max-bs 32 --watchdog-timeout 3000000 --tool-call-parser deepseekv4 --reasoning-parser deepseek-v4 --constrained-json-disable-any-whitespace --enable-metrics --enable-request-time-stats-logging --context-length 140000 --dp-size 4 diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/metrics/sglang_c16_i1024_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/metrics/sglang_c16_i1024_o128_a1.json new file mode 100644 index 0000000..d87202a --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/metrics/sglang_c16_i1024_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 40.23911515297368, + "request_tps": 1.988115287721181, + "input_tps": 2035.8300546264893, + "output_tps": 254.47875682831116, + "total_tps": 2290.3088114548004, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 1052.9030304169282, + "ttft_p95_ms": 0.0, + "ttft_p99_ms": 5166.6259043244645, + "tpot_p50_ms": 49.23821682675089, + "tpot_p95_ms": 0.0, + "tpot_p99_ms": 54.25038263861706, + "e2e_p50_ms": 7454.8611105419695, + "e2e_p95_ms": 0.0, + "e2e_p99_ms": 11466.465564270038, + "itl_p50_ms": 37.7943164203316, + "itl_p95_ms": 63.24442769400776, + "itl_p99_ms": 348.378855206538, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/metrics/sglang_c16_i1024_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/metrics/sglang_c16_i1024_o256_a1.json new file mode 100644 index 0000000..e033509 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/metrics/sglang_c16_i1024_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 57.74114781501703, + "request_tps": 1.3854937601222053, + "input_tps": 1418.7456103651382, + "output_tps": 354.68640259128455, + "total_tps": 1773.4320129564228, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 603.6663899431005, + "ttft_p95_ms": 0.0, + "ttft_p99_ms": 1904.6510633290745, + "tpot_p50_ms": 42.63468751565133, + "tpot_p95_ms": 0.0, + "tpot_p99_ms": 45.214616991625185, + "e2e_p50_ms": 11689.894185517915, + "e2e_p95_ms": 0.0, + "e2e_p99_ms": 12117.081098477356, + "itl_p50_ms": 35.19838349893689, + "itl_p95_ms": 57.88392436224967, + "itl_p99_ms": 250.10306392796338, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/metrics/sglang_c32_i1024_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/metrics/sglang_c32_i1024_o128_a1.json new file mode 100644 index 0000000..2af5079 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/metrics/sglang_c32_i1024_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 50.54340630490333, + "request_tps": 3.165595904534001, + "input_tps": 3241.570206242817, + "output_tps": 405.19627578035215, + "total_tps": 3646.7664820231694, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 2105.71797657758, + "ttft_p95_ms": 0.0, + "ttft_p99_ms": 3751.428431791719, + "tpot_p50_ms": 58.03926556653482, + "tpot_p95_ms": 0.0, + "tpot_p99_ms": 78.91586898356734, + "e2e_p50_ms": 10047.10013512522, + "e2e_p95_ms": 0.0, + "e2e_p99_ms": 11669.661067568231, + "itl_p50_ms": 44.176343944855034, + "itl_p95_ms": 68.84024657774717, + "itl_p99_ms": 453.6125026480295, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/metrics/sglang_c32_i1024_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/metrics/sglang_c32_i1024_o256_a1.json new file mode 100644 index 0000000..e33793a --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/metrics/sglang_c32_i1024_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 70.48552284296602, + "request_tps": 2.2699696837953853, + "input_tps": 2324.4489562064746, + "output_tps": 581.1122390516186, + "total_tps": 2905.5611952580934, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 1804.3314524693415, + "ttft_p95_ms": 0.0, + "ttft_p99_ms": 3630.1654690341093, + "tpot_p50_ms": 47.34475767221667, + "tpot_p95_ms": 0.0, + "tpot_p99_ms": 55.438701826604266, + "e2e_p50_ms": 13931.560064083897, + "e2e_p95_ms": 0.0, + "e2e_p99_ms": 15806.807921521831, + "itl_p50_ms": 38.56221307069063, + "itl_p95_ms": 58.92605511471629, + "itl_p99_ms": 280.3189510037191, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/metrics/sglang_c48_i1024_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/metrics/sglang_c48_i1024_o128_a1.json new file mode 100644 index 0000000..b287d8b --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/metrics/sglang_c48_i1024_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 63.502394546056166, + "request_tps": 3.7793850407630853, + "input_tps": 3870.0902817413994, + "output_tps": 483.7612852176749, + "total_tps": 4353.851566959074, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 3862.513850675896, + "ttft_p95_ms": 0.0, + "ttft_p99_ms": 5555.822190991602, + "tpot_p50_ms": 67.70696836674043, + "tpot_p95_ms": 0.0, + "tpot_p99_ms": 93.07615396471353, + "e2e_p50_ms": 12500.39366458077, + "e2e_p95_ms": 0.0, + "e2e_p99_ms": 16880.625398831908, + "itl_p50_ms": 49.83529110904783, + "itl_p95_ms": 74.06742639141157, + "itl_p99_ms": 259.87648053327575, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/metrics/sglang_c64_i1024_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/metrics/sglang_c64_i1024_o128_a1.json new file mode 100644 index 0000000..3fda2ec --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/metrics/sglang_c64_i1024_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 320, + "failed": 0, + "duration_s": 85.09563249303028, + "request_tps": 3.76047501646115, + "input_tps": 3850.7264168562174, + "output_tps": 481.3408021070272, + "total_tps": 4332.067218963245, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 5960.168475518003, + "ttft_p95_ms": 0.0, + "ttft_p99_ms": 18375.722452029586, + "tpot_p50_ms": 65.91272115795397, + "tpot_p95_ms": 0.0, + "tpot_p99_ms": 185.4429571080667, + "e2e_p50_ms": 14741.17284908425, + "e2e_p95_ms": 0.0, + "e2e_p99_ms": 29969.413060459774, + "itl_p50_ms": 53.08219196740538, + "itl_p95_ms": 75.85310137365013, + "itl_p99_ms": 249.40574671840295, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/metrics/sglang_c80_i1024_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/metrics/sglang_c80_i1024_o128_a1.json new file mode 100644 index 0000000..5f964ea --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/metrics/sglang_c80_i1024_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 400, + "failed": 0, + "duration_s": 107.8278465711046, + "request_tps": 3.7096168820938957, + "input_tps": 3798.647687264149, + "output_tps": 474.83096090801865, + "total_tps": 4273.478648172168, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 7867.146757547744, + "ttft_p95_ms": 0.0, + "ttft_p99_ms": 19524.02482017642, + "tpot_p50_ms": 72.21180715004174, + "tpot_p95_ms": 0.0, + "tpot_p99_ms": 181.08834505535836, + "e2e_p50_ms": 16652.380090556107, + "e2e_p95_ms": 0.0, + "e2e_p99_ms": 31464.33779285988, + "itl_p50_ms": 53.27356734778732, + "itl_p95_ms": 77.84591265954077, + "itl_p99_ms": 331.16745246108627, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/server_cmd.txt b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/server_cmd.txt new file mode 100644 index 0000000..b42f5f5 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-064715/tp4_dp2/server_cmd.txt @@ -0,0 +1 @@ +--host 0.0.0.0 --port 30014 --model-path /models --attention-backend nsa --nsa-prefill klxdsa --nsa-decode klxdsa --trust-remote-code --disable-custom-all-reduce --page-size 64 --mem-fraction-static 0.8 --tensor-parallel-size 4 --disable-shared-experts-fusion --quantization w8a8_int8 --kv-cache-dtype float16 --disable-piecewise-cuda-graph --cuda-graph-max-bs 32 --watchdog-timeout 3000000 --tool-call-parser deepseekv4 --reasoning-parser deepseek-v4 --constrained-json-disable-any-whitespace --enable-metrics --enable-request-time-stats-logging --context-length 140000 --dp-size 2 diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp2_dp4/server_cmd.txt b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp2_dp4/server_cmd.txt new file mode 100644 index 0000000..cddc8dc --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp2_dp4/server_cmd.txt @@ -0,0 +1 @@ +--host 0.0.0.0 --port 30014 --model-path /models --attention-backend nsa --nsa-prefill klxdsa --nsa-decode klxdsa --trust-remote-code --disable-custom-all-reduce --page-size 64 --mem-fraction-static 0.8 --tensor-parallel-size 2 --disable-shared-experts-fusion --quantization w8a8_int8 --kv-cache-dtype float16 --disable-piecewise-cuda-graph --cuda-graph-max-bs 32 --watchdog-timeout 3000000 --tool-call-parser deepseekv4 --reasoning-parser deepseek-v4 --constrained-json-disable-any-whitespace --enable-metrics --enable-request-time-stats-logging --context-length 140000 --dp-size 4 diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i1024_o1024_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i1024_o1024_a1.json new file mode 100644 index 0000000..b0bd6f6 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i1024_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 239.2404650580138, + "request_tps": 0.3343915920770371, + "input_tps": 342.416990286886, + "output_tps": 342.416990286886, + "total_tps": 684.833980573772, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 635.5562334647402, + "ttft_p95_ms": 1903.8669666741043, + "ttft_p99_ms": 1906.7863759305328, + "tpot_p50_ms": 45.73389088913875, + "tpot_p95_ms": 46.81248970833659, + "tpot_p99_ms": 47.236905306400615, + "e2e_p50_ms": 47589.559288928285, + "e2e_p95_ms": 48763.24818015564, + "e2e_p99_ms": 49025.48799880082, + "itl_p50_ms": 38.22040546219796, + "itl_p95_ms": 62.13192297145724, + "itl_p99_ms": 69.20204504393043, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i1024_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i1024_o128_a1.json new file mode 100644 index 0000000..5909184 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i1024_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 41.2688269931823, + "request_tps": 1.9385091806272121, + "input_tps": 1985.0334009622652, + "output_tps": 248.12917512028315, + "total_tps": 2233.1625760825486, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 1839.1637965105474, + "ttft_p95_ms": 5156.00185089279, + "ttft_p99_ms": 5158.878551644739, + "tpot_p50_ms": 47.26517358684809, + "tpot_p95_ms": 54.75502148450271, + "tpot_p99_ms": 55.868444815663054, + "e2e_p50_ms": 7647.177290520631, + "e2e_p95_ms": 11683.607439068146, + "e2e_p99_ms": 11687.237907222006, + "itl_p50_ms": 38.27845409978181, + "itl_p95_ms": 61.400388937909156, + "itl_p99_ms": 83.85505519807339, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i1024_o2048_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i1024_o2048_a1.json new file mode 100644 index 0000000..ca2b03f --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i1024_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 473.66198480315506, + "request_tps": 0.16889681369140588, + "input_tps": 172.95033721999962, + "output_tps": 345.90067443999925, + "total_tps": 518.8510116599988, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 1107.0592959877104, + "ttft_p95_ms": 1914.7031036205587, + "ttft_p99_ms": 1937.422300640028, + "tpot_p50_ms": 45.60600052494192, + "tpot_p95_ms": 46.180353910758, + "tpot_p99_ms": 46.185254772231495, + "e2e_p50_ms": 94571.73220254481, + "e2e_p95_ms": 95557.28795382893, + "e2e_p99_ms": 95606.2740921299, + "itl_p50_ms": 38.73952745925635, + "itl_p95_ms": 62.4737826641649, + "itl_p99_ms": 67.41200149524958, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i1024_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i1024_o256_a1.json new file mode 100644 index 0000000..3ee71a1 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i1024_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 67.59083261387423, + "request_tps": 1.1835924622650464, + "input_tps": 1211.9986813594076, + "output_tps": 302.9996703398519, + "total_tps": 1514.9983516992595, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 1088.1021856330335, + "ttft_p95_ms": 1940.051195200067, + "ttft_p99_ms": 1945.7215382088907, + "tpot_p50_ms": 48.00720819381668, + "tpot_p95_ms": 51.906457407692194, + "tpot_p99_ms": 52.662985878384795, + "e2e_p50_ms": 13323.23444657959, + "e2e_p95_ms": 14546.717717882711, + "e2e_p99_ms": 14597.288672290742, + "itl_p50_ms": 38.109477958641946, + "itl_p95_ms": 62.32210020534694, + "itl_p99_ms": 91.81069683981994, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i1024_o4096_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i1024_o4096_a1.json new file mode 100644 index 0000000..2ad5b3e --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i1024_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 840.8592580729164, + "request_tps": 0.09514077324110604, + "input_tps": 97.42415179889258, + "output_tps": 389.6966071955703, + "total_tps": 487.1207589944629, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 972.442637081258, + "ttft_p95_ms": 5176.270531245973, + "ttft_p99_ms": 5179.259422505274, + "tpot_p50_ms": 40.55560706117929, + "tpot_p95_ms": 40.93768200460964, + "tpot_p99_ms": 40.989130662076434, + "e2e_p50_ms": 167707.99649693072, + "e2e_p95_ms": 171209.72607890144, + "e2e_p99_ms": 171213.10957846232, + "itl_p50_ms": 38.21910894475877, + "itl_p95_ms": 56.424582388717674, + "itl_p99_ms": 61.033719554543495, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i1024_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i1024_o512_a1.json new file mode 100644 index 0000000..6824bd0 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i1024_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 123.35641185496934, + "request_tps": 0.6485272941795385, + "input_tps": 664.0919492398474, + "output_tps": 332.0459746199237, + "total_tps": 996.137923859771, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 1815.1043795514852, + "ttft_p95_ms": 2051.647847087588, + "ttft_p99_ms": 2054.4600446219556, + "tpot_p50_ms": 44.89284182270832, + "tpot_p95_ms": 47.29539701357892, + "tpot_p99_ms": 47.49259979092314, + "e2e_p50_ms": 24692.195138079114, + "e2e_p95_ms": 25352.763180364855, + "e2e_p99_ms": 25618.341889507137, + "itl_p50_ms": 37.906087934970856, + "itl_p95_ms": 61.653098394162946, + "itl_p99_ms": 66.99016668368131, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i131072_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i131072_o128_a1.json new file mode 100644 index 0000000..f021967 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i131072_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 1677.9406901570037, + "request_tps": 0.047677489716584955, + "input_tps": 6249.183932132223, + "output_tps": 6.102718683722874, + "total_tps": 6255.286650815946, + "mean_input_tokens": 131072.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 300625.9035965195, + "ttft_p95_ms": 338951.3012407697, + "ttft_p99_ms": 340203.0271626217, + "tpot_p50_ms": 41.211593476322115, + "tpot_p95_ms": 345.65387451581245, + "tpot_p99_ms": 346.94827124932925, + "e2e_p50_ms": 343158.3466755692, + "e2e_p95_ms": 344526.5441220952, + "e2e_p99_ms": 345839.6717084944, + "itl_p50_ms": 33.60768407583237, + "itl_p95_ms": 58.138947119005024, + "itl_p99_ms": 181.5207198727876, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i16384_o1024_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i16384_o1024_a1.json new file mode 100644 index 0000000..a6a3fc5 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i16384_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 632.0467019469943, + "request_tps": 0.1265729253132138, + "input_tps": 2073.7708083316948, + "output_tps": 129.61067552073092, + "total_tps": 2203.3814838524254, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 80126.8644919619, + "ttft_p95_ms": 106781.9698160165, + "ttft_p99_ms": 107180.66528041381, + "tpot_p50_ms": 38.308278328050854, + "tpot_p95_ms": 42.30265493578899, + "tpot_p99_ms": 45.4902937474118, + "e2e_p50_ms": 120186.66901800316, + "e2e_p95_ms": 148021.29719536752, + "e2e_p99_ms": 148871.62709633118, + "itl_p50_ms": 34.52094295062125, + "itl_p95_ms": 51.694627082906685, + "itl_p99_ms": 59.966907631605864, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i16384_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i16384_o128_a1.json new file mode 100644 index 0000000..ec28e7c --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i16384_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 214.8227549989242, + "request_tps": 0.37240002810875705, + "input_tps": 6101.402060533876, + "output_tps": 47.6672035979209, + "total_tps": 6149.069264131796, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 33368.671427015215, + "ttft_p95_ms": 41899.411271174904, + "ttft_p99_ms": 42125.512621703085, + "tpot_p50_ms": 52.554487114163535, + "tpot_p95_ms": 84.39758658330287, + "tpot_p99_ms": 111.94620711651932, + "e2e_p50_ms": 42495.81566103734, + "e2e_p95_ms": 49210.346029268105, + "e2e_p99_ms": 52561.92248010076, + "itl_p50_ms": 34.52568908687681, + "itl_p95_ms": 52.17893534572764, + "itl_p99_ms": 66.4443611598108, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i16384_o2048_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i16384_o2048_a1.json new file mode 100644 index 0000000..40c3b2d --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i16384_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 1115.2481476869434, + "request_tps": 0.07173291447820136, + "input_tps": 1175.2720708108511, + "output_tps": 146.9090088513564, + "total_tps": 1322.1810796622076, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 136037.42311848328, + "ttft_p95_ms": 181299.1330652847, + "ttft_p99_ms": 181444.37291203765, + "tpot_p50_ms": 37.30771111527574, + "tpot_p95_ms": 39.328867612731, + "tpot_p99_ms": 40.83525040460833, + "e2e_p50_ms": 213226.2396500446, + "e2e_p95_ms": 259703.93927409317, + "e2e_p99_ms": 260337.14725418712, + "itl_p50_ms": 34.49960285797715, + "itl_p95_ms": 51.94193584611639, + "itl_p99_ms": 60.15488663455472, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i16384_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i16384_o256_a1.json new file mode 100644 index 0000000..776927c --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i16384_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 270.44398706802167, + "request_tps": 0.2958098675711304, + "input_tps": 4846.548870285401, + "output_tps": 75.72732609820939, + "total_tps": 4922.27619638361, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 38130.55603799876, + "ttft_p95_ms": 51039.480149233714, + "ttft_p99_ms": 51170.22580615245, + "tpot_p50_ms": 44.34177505135463, + "tpot_p95_ms": 60.19466479852174, + "tpot_p99_ms": 74.01323430483028, + "e2e_p50_ms": 50331.59309858456, + "e2e_p95_ms": 65593.75085806241, + "e2e_p99_ms": 65883.99838647107, + "itl_p50_ms": 34.52300652861595, + "itl_p95_ms": 51.78628868889062, + "itl_p99_ms": 60.336364856921136, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i16384_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i16384_o512_a1.json new file mode 100644 index 0000000..3358e2a --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i16384_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 390.9635538929142, + "request_tps": 0.20462265396204216, + "input_tps": 3352.5375625140987, + "output_tps": 104.76679882856558, + "total_tps": 3457.304361342664, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 52149.89124494605, + "ttft_p95_ms": 69626.60835890565, + "ttft_p99_ms": 69753.87210340472, + "tpot_p50_ms": 40.293584739718284, + "tpot_p95_ms": 48.141768489123194, + "tpot_p99_ms": 54.960999020335215, + "e2e_p50_ms": 73592.7937864326, + "e2e_p95_ms": 93283.15660590306, + "e2e_p99_ms": 93629.28111733867, + "itl_p50_ms": 34.522912465035915, + "itl_p95_ms": 51.489772961940616, + "itl_p99_ms": 59.926832308992736, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i4096_o1024_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i4096_o1024_a1.json new file mode 100644 index 0000000..d5ed5cd --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i4096_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 282.66458361805417, + "request_tps": 0.2830209535839788, + "input_tps": 1159.253825879977, + "output_tps": 289.81345646999426, + "total_tps": 1449.0672823499715, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 4658.318318077363, + "ttft_p95_ms": 10095.44171601301, + "ttft_p99_ms": 12803.387446557112, + "tpot_p50_ms": 44.42285834210774, + "tpot_p95_ms": 49.33753012539997, + "tpot_p99_ms": 79.11847314414067, + "e2e_p50_ms": 49335.88792406954, + "e2e_p95_ms": 58133.938353240934, + "e2e_p99_ms": 87081.42475777536, + "itl_p50_ms": 38.47858391236514, + "itl_p95_ms": 57.68822985701263, + "itl_p99_ms": 62.88705667946487, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i4096_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i4096_o128_a1.json new file mode 100644 index 0000000..c69ff84 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i4096_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 64.04329815506935, + "request_tps": 1.2491549046442667, + "input_tps": 5116.538489422916, + "output_tps": 159.89182779446614, + "total_tps": 5276.430317217383, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 4576.742604491301, + "ttft_p95_ms": 7486.383566202128, + "ttft_p99_ms": 8069.572048275731, + "tpot_p50_ms": 63.25949257419364, + "tpot_p95_ms": 88.51012927140879, + "tpot_p99_ms": 93.78976958789158, + "e2e_p50_ms": 12597.990370937623, + "e2e_p95_ms": 16119.662079052066, + "e2e_p99_ms": 16428.539164408576, + "itl_p50_ms": 38.49607554730028, + "itl_p95_ms": 59.04390099458394, + "itl_p99_ms": 859.1562826023437, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i4096_o2048_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i4096_o2048_a1.json new file mode 100644 index 0000000..33a289b --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i4096_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 534.5376915351953, + "request_tps": 0.14966203743320614, + "input_tps": 613.0157053264123, + "output_tps": 306.50785266320617, + "total_tps": 919.5235579896186, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 5104.571636882611, + "ttft_p95_ms": 13598.845852899827, + "ttft_p99_ms": 15530.32725170254, + "tpot_p50_ms": 42.516595378821116, + "tpot_p95_ms": 48.41014748285291, + "tpot_p99_ms": 80.66643662295293, + "e2e_p50_ms": 90892.7560900338, + "e2e_p95_ms": 106289.35107835554, + "e2e_p99_ms": 171094.83309885245, + "itl_p50_ms": 38.54017308913171, + "itl_p95_ms": 58.33887221524492, + "itl_p99_ms": 64.16293417569251, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i4096_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i4096_o256_a1.json new file mode 100644 index 0000000..350b715 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i4096_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 95.2746979680378, + "request_tps": 0.839677287949398, + "input_tps": 3439.318171440734, + "output_tps": 214.9573857150459, + "total_tps": 3654.27555715578, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 4683.939669514075, + "ttft_p95_ms": 8062.7326244954, + "ttft_p99_ms": 8673.460278843995, + "tpot_p50_ms": 51.88298721149491, + "tpot_p95_ms": 64.92892745028561, + "tpot_p99_ms": 75.71912046924533, + "e2e_p50_ms": 17954.677841393277, + "e2e_p95_ms": 22523.24340294581, + "e2e_p99_ms": 24458.481649178953, + "itl_p50_ms": 38.46407646778971, + "itl_p95_ms": 59.771430026739836, + "itl_p99_ms": 67.52078937599435, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i4096_o4096_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i4096_o4096_a1.json new file mode 100644 index 0000000..488091b --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i4096_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 1029.577203406021, + "request_tps": 0.07770179811222125, + "input_tps": 318.26656506765823, + "output_tps": 318.26656506765823, + "total_tps": 636.5331301353165, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 5483.361787861213, + "ttft_p95_ms": 22374.68960867263, + "ttft_p99_ms": 150920.9240530175, + "tpot_p50_ms": 41.06715219183478, + "tpot_p95_ms": 46.9445546033481, + "tpot_p99_ms": 79.17070015191274, + "e2e_p50_ms": 173329.3942649616, + "e2e_p95_ms": 302973.24356382014, + "e2e_p99_ms": 330154.8701896191, + "itl_p50_ms": 38.36022887844592, + "itl_p95_ms": 57.20986550441011, + "itl_p99_ms": 62.143948464654386, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i4096_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i4096_o512_a1.json new file mode 100644 index 0000000..a36feb2 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i4096_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 160.08099704002962, + "request_tps": 0.4997470123202401, + "input_tps": 2046.9637624637035, + "output_tps": 255.87047030796293, + "total_tps": 2302.8342327716664, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 5474.024421186186, + "ttft_p95_ms": 8615.011906716973, + "ttft_p99_ms": 8951.299454767715, + "tpot_p50_ms": 46.494569923772445, + "tpot_p95_ms": 53.43938896643479, + "tpot_p99_ms": 76.29711307971617, + "e2e_p50_ms": 28335.586922592483, + "e2e_p95_ms": 32704.220330703534, + "e2e_p99_ms": 44918.695579937164, + "itl_p50_ms": 38.48353505600244, + "itl_p95_ms": 59.288766619283706, + "itl_p99_ms": 65.03905458841473, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i65536_o1024_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i65536_o1024_a1.json new file mode 100644 index 0000000..f7ff19e --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i65536_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 1253.2045440499205, + "request_tps": 0.0638363468915201, + "input_tps": 4183.578829882661, + "output_tps": 65.36841921691658, + "total_tps": 4248.947249099578, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 202414.35351443943, + "ttft_p95_ms": 223751.87583953375, + "ttft_p99_ms": 226827.1145781991, + "tpot_p50_ms": 53.94717318803356, + "tpot_p95_ms": 71.33589557350184, + "tpot_p99_ms": 86.74920972760991, + "e2e_p50_ms": 247910.9200029634, + "e2e_p95_ms": 278792.46321712853, + "e2e_p99_ms": 292009.2727964232, + "itl_p50_ms": 36.60361119545996, + "itl_p95_ms": 54.5408145524561, + "itl_p99_ms": 62.74137625936419, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i65536_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i65536_o128_a1.json new file mode 100644 index 0000000..c919b65 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i65536_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 765.7493072859943, + "request_tps": 0.1044728336530135, + "input_tps": 6846.731626283893, + "output_tps": 13.372522707585729, + "total_tps": 6860.104148991479, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 128567.64044612646, + "ttft_p95_ms": 151961.6587853641, + "ttft_p99_ms": 156960.82069837945, + "tpot_p50_ms": 161.96727417372753, + "tpot_p95_ms": 305.1854889790388, + "tpot_p99_ms": 437.706828691014, + "e2e_p50_ms": 146586.26923209522, + "e2e_p95_ms": 179067.64789508888, + "e2e_p99_ms": 186050.76994073341, + "itl_p50_ms": 36.512586986646056, + "itl_p95_ms": 53.31773438956598, + "itl_p99_ms": 61.62503688596189, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i65536_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i65536_o256_a1.json new file mode 100644 index 0000000..6cebf4a --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i65536_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 832.5039105331525, + "request_tps": 0.09609564470245716, + "input_tps": 6297.724171220232, + "output_tps": 24.600485043829032, + "total_tps": 6322.324656264061, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 136335.5990794953, + "ttft_p95_ms": 162983.08086185716, + "ttft_p99_ms": 166663.45600951457, + "tpot_p50_ms": 100.12674616266263, + "tpot_p95_ms": 175.98028106304506, + "tpot_p99_ms": 236.80305566614055, + "e2e_p50_ms": 156982.1368450066, + "e2e_p95_ms": 189405.96394817112, + "e2e_p99_ms": 206104.3898640247, + "itl_p50_ms": 36.563907051458955, + "itl_p95_ms": 53.42142162844535, + "itl_p99_ms": 62.24962119013072, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i65536_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i65536_o512_a1.json new file mode 100644 index 0000000..64630a7 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c16_i65536_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 973.7963511580601, + "request_tps": 0.08215270051573127, + "input_tps": 5383.959380998965, + "output_tps": 42.06218266405441, + "total_tps": 5426.021563663019, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 153779.00782134384, + "ttft_p95_ms": 181627.71489518927, + "ttft_p99_ms": 183322.4648270011, + "tpot_p50_ms": 69.34998192962937, + "tpot_p95_ms": 106.79924657778838, + "tpot_p99_ms": 136.58104994009412, + "e2e_p50_ms": 189005.7152614463, + "e2e_p95_ms": 234631.01633958286, + "e2e_p99_ms": 236243.1676932983, + "itl_p50_ms": 36.63114202208817, + "itl_p95_ms": 53.81391057744622, + "itl_p99_ms": 62.73157489486038, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i131072_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i131072_o128_a1.json new file mode 100644 index 0000000..1e8db1b --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i131072_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 226.54693996487185, + "request_tps": 0.022070481290876385, + "input_tps": 2892.8221237577495, + "output_tps": 2.8250216052321773, + "total_tps": 2895.647145362982, + "mean_input_tokens": 131072.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 40983.879873063415, + "ttft_p95_ms": 41427.96418489888, + "ttft_p99_ms": 41483.68903929368, + "tpot_p50_ms": 33.28039894415814, + "tpot_p95_ms": 33.693099089744166, + "tpot_p99_ms": 33.698522558318466, + "e2e_p50_ms": 45259.45955491625, + "e2e_p95_ms": 45639.66128444299, + "e2e_p99_ms": 45701.79939834401, + "itl_p50_ms": 30.720365000888705, + "itl_p95_ms": 55.317965033464134, + "itl_p99_ms": 62.53650498110801, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i16384_o1024_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i16384_o1024_a1.json new file mode 100644 index 0000000..4aa8bea --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i16384_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 181.57242772518657, + "request_tps": 0.02753722061571814, + "input_tps": 451.169822567926, + "output_tps": 28.198113910495376, + "total_tps": 479.3679364784214, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 3940.7668448984623, + "ttft_p95_ms": 4193.481011828408, + "ttft_p99_ms": 4228.460800005123, + "tpot_p50_ms": 31.63807294540879, + "tpot_p95_ms": 31.730153837458815, + "tpot_p99_ms": 31.734481836830287, + "e2e_p50_ms": 36320.72346494533, + "e2e_p95_ms": 36463.93566392362, + "e2e_p99_ms": 36480.87247610092, + "itl_p50_ms": 29.840212780982256, + "itl_p95_ms": 46.52133162599058, + "itl_p99_ms": 54.09107577055689, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i16384_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i16384_o128_a1.json new file mode 100644 index 0000000..ebbe95c --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i16384_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 39.83333700802177, + "request_tps": 0.12552300097260452, + "input_tps": 2056.5688479351525, + "output_tps": 16.06694412449338, + "total_tps": 2072.6357920596456, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 3952.2010260261595, + "ttft_p95_ms": 4079.302547173574, + "ttft_p99_ms": 4102.609633402899, + "tpot_p50_ms": 31.602383598567936, + "tpot_p95_ms": 31.640485883140425, + "tpot_p99_ms": 31.642765031291507, + "e2e_p50_ms": 7969.095474109054, + "e2e_p95_ms": 8074.662221223116, + "e2e_p99_ms": 8093.300537876784, + "itl_p50_ms": 29.8505169339478, + "itl_p95_ms": 49.705708399415016, + "itl_p99_ms": 51.2071222346276, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i16384_o2048_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i16384_o2048_a1.json new file mode 100644 index 0000000..a323701 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i16384_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 342.39226463483647, + "request_tps": 0.014603133646528293, + "input_tps": 239.25774166471956, + "output_tps": 29.907217708089945, + "total_tps": 269.1649593728095, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 3971.6174241621047, + "ttft_p95_ms": 4094.661511434242, + "ttft_p99_ms": 4115.855502327904, + "tpot_p50_ms": 31.558119324293, + "tpot_p95_ms": 31.620720006947725, + "tpot_p99_ms": 31.632591595734013, + "e2e_p50_ms": 68487.32287203893, + "e2e_p95_ms": 68711.56300739385, + "e2e_p99_ms": 68739.65807267465, + "itl_p50_ms": 29.78591388091445, + "itl_p95_ms": 45.330130099318914, + "itl_p99_ms": 55.03684663679451, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i16384_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i16384_o256_a1.json new file mode 100644 index 0000000..be6b00e --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i16384_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 60.537773119984195, + "request_tps": 0.08259306119652168, + "input_tps": 1353.2047146438113, + "output_tps": 21.14382366630955, + "total_tps": 1374.3485383101208, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 3945.7851310726255, + "ttft_p95_ms": 4093.810023879632, + "ttft_p99_ms": 4111.218916652724, + "tpot_p50_ms": 31.821994341033346, + "tpot_p95_ms": 32.01389185737307, + "tpot_p99_ms": 32.03676525156433, + "e2e_p50_ms": 12080.163977108896, + "e2e_p95_ms": 12253.15336487256, + "e2e_p99_ms": 12279.754239274189, + "itl_p50_ms": 29.860365903005004, + "itl_p95_ms": 48.98134677205235, + "itl_p99_ms": 55.7887158729136, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i16384_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i16384_o512_a1.json new file mode 100644 index 0000000..9c1662c --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i16384_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 100.54988616588525, + "request_tps": 0.04972656052291394, + "input_tps": 814.719967607422, + "output_tps": 25.459998987731936, + "total_tps": 840.1799665951539, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 3932.124942075461, + "ttft_p95_ms": 4127.591159660369, + "ttft_p99_ms": 4166.129624564201, + "tpot_p50_ms": 31.635086914185443, + "tpot_p95_ms": 31.776042928674975, + "tpot_p99_ms": 31.788027078066378, + "e2e_p50_ms": 20141.837269999087, + "e2e_p95_ms": 20233.498825039715, + "e2e_p99_ms": 20244.73103923723, + "itl_p50_ms": 29.790908098220825, + "itl_p95_ms": 50.02487420570105, + "itl_p99_ms": 54.69925904646516, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i65536_o1024_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i65536_o1024_a1.json new file mode 100644 index 0000000..18a612e --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i65536_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 256.18154512299225, + "request_tps": 0.019517409021791596, + "input_tps": 1279.092917652134, + "output_tps": 19.985826838314594, + "total_tps": 1299.0787444904488, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 17658.26961491257, + "ttft_p95_ms": 18013.91982375644, + "ttft_p99_ms": 18047.10275827907, + "tpot_p50_ms": 32.7376385434701, + "tpot_p95_ms": 32.79802809666955, + "tpot_p99_ms": 32.798152351014515, + "e2e_p50_ms": 51210.01679683104, + "e2e_p95_ms": 51457.70765142515, + "e2e_p99_ms": 51481.52730548754, + "itl_p50_ms": 30.361955054104328, + "itl_p95_ms": 53.69571931660175, + "itl_p99_ms": 60.88819689117371, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i65536_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i65536_o128_a1.json new file mode 100644 index 0000000..b4118f8 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i65536_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 109.37776633584872, + "request_tps": 0.045713129528055126, + "input_tps": 2995.8556567506207, + "output_tps": 5.851280579591056, + "total_tps": 3001.706937330212, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 17729.769844096154, + "ttft_p95_ms": 17954.093989543617, + "ttft_p99_ms": 17964.415912255645, + "tpot_p50_ms": 32.0961656301891, + "tpot_p95_ms": 32.58936951105578, + "tpot_p99_ms": 32.640932479609305, + "e2e_p50_ms": 21835.877286968753, + "e2e_p95_ms": 22044.52510732226, + "e2e_p99_ms": 22048.520961580798, + "itl_p50_ms": 30.422291019931436, + "itl_p95_ms": 47.328294953331344, + "itl_p99_ms": 56.53701915405691, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i65536_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i65536_o256_a1.json new file mode 100644 index 0000000..9b95d7a --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i65536_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 129.80575013905764, + "request_tps": 0.038519094837044014, + "input_tps": 2524.3873992405165, + "output_tps": 9.860888278283268, + "total_tps": 2534.2482875187998, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 17821.7026530765, + "ttft_p95_ms": 17941.041450900957, + "ttft_p99_ms": 17956.14996048622, + "tpot_p50_ms": 32.05825664723913, + "tpot_p95_ms": 32.285422013122954, + "tpot_p99_ms": 32.28693708328201, + "e2e_p50_ms": 26052.55355197005, + "e2e_p95_ms": 26127.57891532965, + "e2e_p99_ms": 26133.34180940874, + "itl_p50_ms": 30.41153005324304, + "itl_p95_ms": 46.34950009640306, + "itl_p99_ms": 56.67572288773954, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i65536_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i65536_o512_a1.json new file mode 100644 index 0000000..279a883 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c1_i65536_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 170.8164885980077, + "request_tps": 0.029271178918604212, + "input_tps": 1918.3159816096456, + "output_tps": 14.986843606325357, + "total_tps": 1933.302825215971, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 17657.30693913065, + "ttft_p95_ms": 17947.263365006074, + "ttft_p99_ms": 17974.353000214323, + "tpot_p50_ms": 32.16144313886325, + "tpot_p95_ms": 32.28488007000849, + "tpot_p99_ms": 32.28917933327971, + "e2e_p50_ms": 34091.804383089766, + "e2e_p95_ms": 34388.406378310174, + "e2e_p99_ms": 34406.406796555966, + "itl_p50_ms": 30.36797302775085, + "itl_p95_ms": 46.908468334004276, + "itl_p99_ms": 55.469688819721355, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c32_i1024_o1024_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c32_i1024_o1024_a1.json new file mode 100644 index 0000000..248f358 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c32_i1024_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 283.1469709521625, + "request_tps": 0.5650775618822773, + "input_tps": 578.639423367452, + "output_tps": 578.639423367452, + "total_tps": 1157.278846734904, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 1039.5908689824864, + "ttft_p95_ms": 3708.7073334376323, + "ttft_p99_ms": 3736.5557909896597, + "tpot_p50_ms": 53.18489079777869, + "tpot_p95_ms": 56.531733996098275, + "tpot_p99_ms": 56.90662554211692, + "e2e_p50_ms": 55562.747771968134, + "e2e_p95_ms": 58622.54847800359, + "e2e_p99_ms": 58864.768436276354, + "itl_p50_ms": 44.60177046712488, + "itl_p95_ms": 68.54137406917289, + "itl_p99_ms": 104.98853123513977, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c32_i1024_o2048_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c32_i1024_o2048_a1.json new file mode 100644 index 0000000..6b099ab --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c32_i1024_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 562.1656430249568, + "request_tps": 0.28461362231077675, + "input_tps": 291.4443492462354, + "output_tps": 582.8886984924708, + "total_tps": 874.3330477387061, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 1720.5767229897901, + "ttft_p95_ms": 3765.552905423101, + "ttft_p99_ms": 3778.2479317323305, + "tpot_p50_ms": 53.25908665974094, + "tpot_p95_ms": 55.20441076613287, + "tpot_p99_ms": 55.39557766710268, + "e2e_p50_ms": 109979.57844298799, + "e2e_p95_ms": 114323.36549273459, + "e2e_p99_ms": 114707.30229713488, + "itl_p50_ms": 46.27890046685934, + "itl_p95_ms": 69.01592136127874, + "itl_p99_ms": 77.31038637226447, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c32_i1024_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c32_i1024_o256_a1.json new file mode 100644 index 0000000..42543c1 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c32_i1024_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 85.51100934692658, + "request_tps": 1.8711040978462112, + "input_tps": 1916.0105961945203, + "output_tps": 479.0026490486301, + "total_tps": 2395.0132452431503, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 1161.2384765176103, + "ttft_p95_ms": 3756.2352412380274, + "ttft_p99_ms": 3780.965418519918, + "tpot_p50_ms": 61.98459189412568, + "tpot_p95_ms": 66.44078092480663, + "tpot_p99_ms": 66.8051291848259, + "e2e_p50_ms": 16901.625977130607, + "e2e_p95_ms": 18837.460162944622, + "e2e_p99_ms": 18985.462763826363, + "itl_p50_ms": 44.66060339473188, + "itl_p95_ms": 68.57059714384377, + "itl_p99_ms": 449.1709415311925, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c32_i1024_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c32_i1024_o512_a1.json new file mode 100644 index 0000000..bb5d004 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c32_i1024_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 151.8317316111643, + "request_tps": 1.0537981639421354, + "input_tps": 1079.0893198767467, + "output_tps": 539.5446599383733, + "total_tps": 1618.63397981512, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 1056.6700529307127, + "ttft_p95_ms": 3730.0430950359428, + "ttft_p99_ms": 3803.3011273038574, + "tpot_p50_ms": 55.7243038511538, + "tpot_p95_ms": 60.350505806735306, + "tpot_p99_ms": 61.641385694806246, + "e2e_p50_ms": 29571.674760896713, + "e2e_p95_ms": 32270.35216266522, + "e2e_p99_ms": 32279.17558920337, + "itl_p50_ms": 45.009728986769915, + "itl_p95_ms": 67.9298322298564, + "itl_p99_ms": 236.92634512903186, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c48_i1024_o1024_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c48_i1024_o1024_a1.json new file mode 100644 index 0000000..d979b73 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c48_i1024_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 386.91459054895677, + "request_tps": 0.6202919348672961, + "input_tps": 635.1789413041112, + "output_tps": 635.1789413041112, + "total_tps": 1270.3578826082223, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 5370.0150235090405, + "ttft_p95_ms": 15133.30938139696, + "ttft_p99_ms": 32185.267814283718, + "tpot_p50_ms": 60.0038268968676, + "tpot_p95_ms": 91.36051824342555, + "tpot_p99_ms": 103.14346713503731, + "e2e_p50_ms": 65575.55996044539, + "e2e_p95_ms": 99691.6358396411, + "e2e_p99_ms": 112446.33609429697, + "itl_p50_ms": 51.48644756991416, + "itl_p95_ms": 73.74423729488626, + "itl_p99_ms": 83.97575234295799, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c48_i1024_o2048_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c48_i1024_o2048_a1.json new file mode 100644 index 0000000..99e22ea --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c48_i1024_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 785.6881854759995, + "request_tps": 0.3054646925288802, + "input_tps": 312.7958451495733, + "output_tps": 625.5916902991466, + "total_tps": 938.3875354487201, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 5525.519606540911, + "ttft_p95_ms": 38321.421813906585, + "ttft_p99_ms": 61896.90983353647, + "tpot_p50_ms": 60.786382354451064, + "tpot_p95_ms": 94.29399440951978, + "tpot_p99_ms": 106.96621055514711, + "e2e_p50_ms": 128179.90333400667, + "e2e_p95_ms": 211186.33220964108, + "e2e_p99_ms": 235519.09909247872, + "itl_p50_ms": 52.96039255335927, + "itl_p95_ms": 74.77014367468655, + "itl_p99_ms": 93.00403241533787, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c48_i1024_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c48_i1024_o256_a1.json new file mode 100644 index 0000000..ec3349e --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c48_i1024_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 102.57387904892676, + "request_tps": 2.3397769707580456, + "input_tps": 2395.9316180562387, + "output_tps": 598.9829045140597, + "total_tps": 2994.9145225702982, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 4857.287564082071, + "ttft_p95_ms": 5687.151023850311, + "ttft_p99_ms": 5947.1146802348085, + "tpot_p50_ms": 60.30845692353871, + "tpot_p95_ms": 72.36586667291418, + "tpot_p99_ms": 78.81875480012964, + "e2e_p50_ms": 20432.19355738256, + "e2e_p95_ms": 21489.17214354733, + "e2e_p99_ms": 22516.328958119266, + "itl_p50_ms": 51.60140746738762, + "itl_p95_ms": 73.69143945397809, + "itl_p99_ms": 89.00051706237718, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c48_i1024_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c48_i1024_o512_a1.json new file mode 100644 index 0000000..9350fa6 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c48_i1024_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 194.6086190820206, + "request_tps": 1.2332444530570794, + "input_tps": 1262.8423199304493, + "output_tps": 631.4211599652247, + "total_tps": 1894.263479895674, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 5219.216491095722, + "ttft_p95_ms": 5779.679291823413, + "ttft_p99_ms": 6389.283453426324, + "tpot_p50_ms": 58.08764942142558, + "tpot_p95_ms": 62.69429164324429, + "tpot_p99_ms": 91.06430042089896, + "e2e_p50_ms": 34896.657894947566, + "e2e_p95_ms": 35910.429730673786, + "e2e_p99_ms": 49593.8836463075, + "itl_p50_ms": 51.571647985838354, + "itl_p95_ms": 73.75171015737578, + "itl_p99_ms": 77.6099856174551, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i1024_o4096_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i1024_o4096_a1.json new file mode 100644 index 0000000..af5a11a --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i1024_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 750.264118807856, + "request_tps": 0.0533145581632754, + "input_tps": 54.59410755919401, + "output_tps": 218.37643023677603, + "total_tps": 272.97053779597, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 457.8002068446949, + "ttft_p95_ms": 1004.7794182202779, + "ttft_p99_ms": 1006.1946984101087, + "tpot_p50_ms": 36.49401828900812, + "tpot_p95_ms": 36.67125693285014, + "tpot_p99_ms": 36.679555167656694, + "e2e_p50_ms": 150253.00148909446, + "e2e_p95_ms": 150511.7817995837, + "e2e_p99_ms": 150794.05072218738, + "itl_p50_ms": 34.35514506418258, + "itl_p95_ms": 53.22418445721268, + "itl_p99_ms": 59.324696937110275, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i131072_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i131072_o128_a1.json new file mode 100644 index 0000000..2723e11 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i131072_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 818.1570529809687, + "request_tps": 0.04889036873086816, + "input_tps": 6408.158410292352, + "output_tps": 6.257967197551125, + "total_tps": 6414.4163774899025, + "mean_input_tokens": 131072.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 129584.32166394778, + "ttft_p95_ms": 167181.3310468453, + "ttft_p99_ms": 168040.67650724668, + "tpot_p50_ms": 36.84465442857463, + "tpot_p95_ms": 343.3015784548456, + "tpot_p99_ms": 345.92269529082523, + "e2e_p50_ms": 171526.36588714086, + "e2e_p95_ms": 173146.40125581063, + "e2e_p99_ms": 173291.67424835498, + "itl_p50_ms": 33.540375996381044, + "itl_p95_ms": 58.00506006926298, + "itl_p99_ms": 276.8040000228211, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i16384_o1024_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i16384_o1024_a1.json new file mode 100644 index 0000000..e8fdb52 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i16384_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 335.39274771884084, + "request_tps": 0.11926316317826863, + "input_tps": 1954.0076655127532, + "output_tps": 122.12547909454707, + "total_tps": 2076.1331446073004, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 11495.135091943666, + "ttft_p95_ms": 54293.259906279854, + "ttft_p99_ms": 54475.75780400075, + "tpot_p50_ms": 41.142632146611746, + "tpot_p95_ms": 44.4825864581911, + "tpot_p99_ms": 48.84856583131568, + "e2e_p50_ms": 49175.88768794667, + "e2e_p95_ms": 97491.41289745457, + "e2e_p99_ms": 102364.25640797475, + "itl_p50_ms": 34.5357779879123, + "itl_p95_ms": 52.090133994352065, + "itl_p99_ms": 60.2705589751713, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i16384_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i16384_o128_a1.json new file mode 100644 index 0000000..36c0cfa --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i16384_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 109.44023031182587, + "request_tps": 0.36549630685195744, + "input_tps": 5988.291491462471, + "output_tps": 46.78352727705055, + "total_tps": 6035.075018739521, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 11529.161532991566, + "ttft_p95_ms": 21725.308789138217, + "ttft_p99_ms": 21941.029288116843, + "tpot_p50_ms": 76.67213628768891, + "tpot_p95_ms": 101.16250104450535, + "tpot_p99_ms": 138.8094991766004, + "e2e_p50_ms": 16583.38130300399, + "e2e_p95_ms": 32255.35403857939, + "e2e_p99_ms": 37416.11094208899, + "itl_p50_ms": 34.63044541422278, + "itl_p95_ms": 54.926601552870125, + "itl_p99_ms": 1767.2955490718596, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i16384_o2048_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i16384_o2048_a1.json new file mode 100644 index 0000000..587dcdb --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i16384_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 584.8935852469876, + "request_tps": 0.06838850862607578, + "input_tps": 1120.4773253296255, + "output_tps": 140.0596656662032, + "total_tps": 1260.5369909958285, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 11495.196483563632, + "ttft_p95_ms": 91582.56554905092, + "ttft_p99_ms": 91693.05469256826, + "tpot_p50_ms": 38.41638142817654, + "tpot_p95_ms": 41.38107354754197, + "tpot_p99_ms": 42.587399352946996, + "e2e_p50_ms": 89565.8936999971, + "e2e_p95_ms": 172048.6941325711, + "e2e_p99_ms": 176627.42560487473, + "itl_p50_ms": 34.59935949649662, + "itl_p95_ms": 51.98211170500145, + "itl_p99_ms": 60.23820979986343, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i16384_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i16384_o256_a1.json new file mode 100644 index 0000000..250cade --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i16384_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 141.670446844073, + "request_tps": 0.28234540718308926, + "input_tps": 4625.947151287734, + "output_tps": 72.28042423887085, + "total_tps": 4698.2275755266055, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 11554.360230453312, + "ttft_p95_ms": 26391.11408374738, + "ttft_p99_ms": 26587.14834881015, + "tpot_p50_ms": 56.31336195065695, + "tpot_p95_ms": 68.25714148759073, + "tpot_p99_ms": 87.6372731319245, + "e2e_p50_ms": 21265.938312513754, + "e2e_p95_ms": 41577.943935687654, + "e2e_p99_ms": 46795.674697651986, + "itl_p50_ms": 34.58584542386234, + "itl_p95_ms": 52.59093521162865, + "itl_p99_ms": 61.120713129639775, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i16384_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i16384_o512_a1.json new file mode 100644 index 0000000..b2bdbc3 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i16384_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 206.38675206596963, + "request_tps": 0.19381088950522546, + "input_tps": 3175.397613653614, + "output_tps": 99.23117542667543, + "total_tps": 3274.628789080289, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 11589.022769010626, + "ttft_p95_ms": 35687.3680736986, + "ttft_p99_ms": 35902.96996307094, + "tpot_p50_ms": 46.31837186974942, + "tpot_p95_ms": 52.744000344951814, + "tpot_p99_ms": 61.98560190418053, + "e2e_p50_ms": 30626.526765991002, + "e2e_p95_ms": 60265.1604500832, + "e2e_p99_ms": 65432.796396943726, + "itl_p50_ms": 34.61834159679711, + "itl_p95_ms": 51.886969921179116, + "itl_p99_ms": 60.28145674150437, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i4096_o1024_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i4096_o1024_a1.json new file mode 100644 index 0000000..83466fc --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i4096_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 207.04861099808477, + "request_tps": 0.1931913467430603, + "input_tps": 791.311756259575, + "output_tps": 197.82793906489374, + "total_tps": 989.1396953244687, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 1074.1113065741956, + "ttft_p95_ms": 3878.6458508693613, + "ttft_p99_ms": 3916.904898113571, + "tpot_p50_ms": 39.24930339234399, + "tpot_p95_ms": 40.05846651063872, + "tpot_p99_ms": 40.82051611248493, + "e2e_p50_ms": 41186.332597979344, + "e2e_p95_ms": 43346.47120544687, + "e2e_p99_ms": 43701.94057547487, + "itl_p50_ms": 34.467250341549516, + "itl_p95_ms": 54.143763531465076, + "itl_p99_ms": 59.957590696867555, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i4096_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i4096_o128_a1.json new file mode 100644 index 0000000..198e7c9 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i4096_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 42.330964882159606, + "request_tps": 0.9449347566574843, + "input_tps": 3870.4527632690556, + "output_tps": 120.95164885215799, + "total_tps": 3991.4044121212137, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 2789.9403674528003, + "ttft_p95_ms": 3850.861389178317, + "ttft_p99_ms": 3857.0726825506426, + "tpot_p50_ms": 44.318099374200884, + "tpot_p95_ms": 52.61084187132357, + "tpot_p99_ms": 56.91049005684083, + "e2e_p50_ms": 8397.533455048688, + "e2e_p95_ms": 10301.826596679166, + "e2e_p99_ms": 10350.095471963286, + "itl_p50_ms": 34.42738042213023, + "itl_p95_ms": 54.41078649600968, + "itl_p99_ms": 108.43641549814492, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i4096_o2048_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i4096_o2048_a1.json new file mode 100644 index 0000000..42f0ffd --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i4096_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 394.49152314290404, + "request_tps": 0.10139634859913087, + "input_tps": 415.31944386204003, + "output_tps": 207.65972193102002, + "total_tps": 622.97916579306, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 1034.7788120852783, + "ttft_p95_ms": 3756.3848297810177, + "ttft_p99_ms": 3805.5001059989445, + "tpot_p50_ms": 37.88820180411141, + "tpot_p95_ms": 38.25126371524308, + "tpot_p99_ms": 38.27709748982381, + "e2e_p50_ms": 78674.81265543029, + "e2e_p95_ms": 80982.15972669423, + "e2e_p99_ms": 80984.67170314165, + "itl_p50_ms": 34.52226798981428, + "itl_p95_ms": 53.908159665297724, + "itl_p99_ms": 59.665183436591185, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i4096_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i4096_o256_a1.json new file mode 100644 index 0000000..4d7b8a8 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i4096_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 65.8774013610091, + "request_tps": 0.6071884921628803, + "input_tps": 2487.0440638991577, + "output_tps": 155.44025399369735, + "total_tps": 2642.4843178928554, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 1971.202465472743, + "ttft_p95_ms": 3824.870833219029, + "ttft_p99_ms": 3841.111220801249, + "tpot_p50_ms": 43.596625505654394, + "tpot_p95_ms": 48.1240840446131, + "tpot_p99_ms": 49.99360711309218, + "e2e_p50_ms": 13162.176689947955, + "e2e_p95_ms": 14974.834044638555, + "e2e_p99_ms": 14991.584688341245, + "itl_p50_ms": 34.47401651646942, + "itl_p95_ms": 54.485111916437745, + "itl_p99_ms": 60.43242933228612, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i4096_o4096_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i4096_o4096_a1.json new file mode 100644 index 0000000..fb82ac9 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i4096_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 762.6022247918881, + "request_tps": 0.052451984402374224, + "input_tps": 214.84332811212482, + "output_tps": 214.84332811212482, + "total_tps": 429.68665622424965, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 1038.2880391553044, + "ttft_p95_ms": 3893.24660198763, + "ttft_p99_ms": 3900.2761626965366, + "tpot_p50_ms": 36.76516193150564, + "tpot_p95_ms": 37.05582632169512, + "tpot_p99_ms": 37.09407625586006, + "e2e_p50_ms": 152243.55608585756, + "e2e_p95_ms": 154084.0833232156, + "e2e_p99_ms": 154084.53153002774, + "itl_p50_ms": 34.463146002963185, + "itl_p95_ms": 51.49366667028517, + "itl_p99_ms": 59.961400236934416, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i4096_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i4096_o512_a1.json new file mode 100644 index 0000000..b22f964 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i4096_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 112.8795186879579, + "request_tps": 0.3543601218798184, + "input_tps": 1451.4590592197362, + "output_tps": 181.43238240246703, + "total_tps": 1632.8914416222033, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 1941.342479432933, + "ttft_p95_ms": 3813.8611613423564, + "ttft_p99_ms": 3815.337903073523, + "tpot_p50_ms": 40.17744154405986, + "tpot_p95_ms": 42.320974605816815, + "tpot_p99_ms": 43.39261491248045, + "e2e_p50_ms": 22431.270663044415, + "e2e_p95_ms": 24335.910513892304, + "e2e_p99_ms": 24337.53361347597, + "itl_p50_ms": 34.43968354258686, + "itl_p95_ms": 54.208886146079735, + "itl_p99_ms": 59.69965270254761, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i65536_o1024_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i65536_o1024_a1.json new file mode 100644 index 0000000..213f20b --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i65536_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 626.4431783319451, + "request_tps": 0.06385255899267604, + "input_tps": 4184.641306144017, + "output_tps": 65.38502040850027, + "total_tps": 4250.026326552517, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 51137.51745200716, + "ttft_p95_ms": 96054.72948674572, + "ttft_p99_ms": 112415.82493769703, + "tpot_p50_ms": 70.99086488608475, + "tpot_p95_ms": 93.18777094192289, + "tpot_p99_ms": 94.39113497935946, + "e2e_p50_ms": 112995.05897297058, + "e2e_p95_ms": 168512.40640780888, + "e2e_p99_ms": 178801.59651178867, + "itl_p50_ms": 36.47732094395906, + "itl_p95_ms": 61.783290666062385, + "itl_p99_ms": 68.58176962938157, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i65536_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i65536_o128_a1.json new file mode 100644 index 0000000..f05750d --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i65536_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 386.69948341907, + "request_tps": 0.10343949685769714, + "input_tps": 6779.01086606604, + "output_tps": 13.240255597785234, + "total_tps": 6792.251121663825, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 52697.72171450313, + "ttft_p95_ms": 77315.9501316608, + "ttft_p99_ms": 77702.88953800453, + "tpot_p50_ms": 161.86251543282611, + "tpot_p95_ms": 301.6361779644242, + "tpot_p99_ms": 390.316338291553, + "e2e_p50_ms": 73942.72513350006, + "e2e_p95_ms": 98705.0266831764, + "e2e_p99_ms": 108562.47204806423, + "itl_p50_ms": 36.458888091146946, + "itl_p95_ms": 54.5758139109239, + "itl_p99_ms": 62.127575105987496, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i65536_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i65536_o256_a1.json new file mode 100644 index 0000000..cc1b59b --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i65536_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 419.1958255460486, + "request_tps": 0.09542079754228375, + "input_tps": 6253.497387731108, + "output_tps": 24.42772417082464, + "total_tps": 6277.925111901933, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 52760.39652095642, + "ttft_p95_ms": 81073.16757803783, + "ttft_p99_ms": 81568.98809597129, + "tpot_p50_ms": 99.82632672934629, + "tpot_p95_ms": 169.0839746892087, + "tpot_p99_ms": 210.72348307306856, + "e2e_p50_ms": 80142.28945260402, + "e2e_p95_ms": 107118.48287502071, + "e2e_p99_ms": 117810.70945826358, + "itl_p50_ms": 36.5109599661082, + "itl_p95_ms": 53.218919737264514, + "itl_p99_ms": 61.9426407897845, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i65536_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i65536_o512_a1.json new file mode 100644 index 0000000..b79cd28 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/metrics/sglang_c8_i65536_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 487.3736931588501, + "request_tps": 0.08207254630578259, + "input_tps": 5378.706394695768, + "output_tps": 42.021143708560686, + "total_tps": 5420.727538404329, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 52526.89733938314, + "ttft_p95_ms": 84162.39211051725, + "ttft_p99_ms": 89592.60615824256, + "tpot_p50_ms": 77.38403851756873, + "tpot_p95_ms": 114.1166405163389, + "tpot_p99_ms": 167.8516751349313, + "e2e_p50_ms": 79288.78492198419, + "e2e_p95_ms": 137056.76206527278, + "e2e_p99_ms": 142416.35865559801, + "itl_p50_ms": 36.64115187712014, + "itl_p95_ms": 55.09973680600524, + "itl_p99_ms": 62.820788719691336, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/server_cmd.txt b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/server_cmd.txt new file mode 100644 index 0000000..b42f5f5 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp4_dp2/server_cmd.txt @@ -0,0 +1 @@ +--host 0.0.0.0 --port 30014 --model-path /models --attention-backend nsa --nsa-prefill klxdsa --nsa-decode klxdsa --trust-remote-code --disable-custom-all-reduce --page-size 64 --mem-fraction-static 0.8 --tensor-parallel-size 4 --disable-shared-experts-fusion --quantization w8a8_int8 --kv-cache-dtype float16 --disable-piecewise-cuda-graph --cuda-graph-max-bs 32 --watchdog-timeout 3000000 --tool-call-parser deepseekv4 --reasoning-parser deepseek-v4 --constrained-json-disable-any-whitespace --enable-metrics --enable-request-time-stats-logging --context-length 140000 --dp-size 2 diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i1024_o1024_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i1024_o1024_a1.json new file mode 100644 index 0000000..bfc4d18 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i1024_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 195.86190408095717, + "request_tps": 0.40845104807585736, + "input_tps": 418.25387322967794, + "output_tps": 418.25387322967794, + "total_tps": 836.5077464593559, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 2806.6968551138416, + "ttft_p95_ms": 2903.0821086023934, + "ttft_p99_ms": 2903.6138558760285, + "tpot_p50_ms": 35.50938803068511, + "tpot_p95_ms": 36.587063905092016, + "tpot_p99_ms": 36.846195016498285, + "e2e_p50_ms": 39140.523318434134, + "e2e_p95_ms": 39359.496829425916, + "e2e_p99_ms": 39362.11585983634, + "itl_p50_ms": 33.30066951457411, + "itl_p95_ms": 52.0355909015052, + "itl_p99_ms": 61.44081162754446, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i1024_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i1024_o128_a1.json new file mode 100644 index 0000000..e427a70 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i1024_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 40.91812721895985, + "request_tps": 1.9551236930250107, + "input_tps": 2002.046661657611, + "output_tps": 250.25583270720136, + "total_tps": 2252.3024943648124, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 2898.4853614820167, + "ttft_p95_ms": 5916.232369153295, + "ttft_p99_ms": 5917.164536933415, + "tpot_p50_ms": 36.42805760238671, + "tpot_p95_ms": 45.71566509713573, + "tpot_p99_ms": 52.58390468757908, + "e2e_p50_ms": 7513.254255056381, + "e2e_p95_ms": 10505.285932659172, + "e2e_p99_ms": 10506.672543857712, + "itl_p50_ms": 33.98504899814725, + "itl_p95_ms": 55.32407406717539, + "itl_p99_ms": 61.596665140241385, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i1024_o2048_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i1024_o2048_a1.json new file mode 100644 index 0000000..f314ef1 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i1024_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 375.86731938296, + "request_tps": 0.21284106352031737, + "input_tps": 217.949249044805, + "output_tps": 435.89849808961, + "total_tps": 653.8477471344149, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 1012.9853350808844, + "ttft_p95_ms": 1060.0732530583628, + "ttft_p99_ms": 1060.6082381820306, + "tpot_p50_ms": 36.26880324916558, + "tpot_p95_ms": 36.36031736202677, + "tpot_p99_ms": 36.50945045857321, + "e2e_p50_ms": 75173.97456045728, + "e2e_p95_ms": 75301.6543291742, + "e2e_p99_ms": 75354.01330835419, + "itl_p50_ms": 34.3891941010952, + "itl_p95_ms": 56.740176025778055, + "itl_p99_ms": 61.57854507677257, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i1024_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i1024_o256_a1.json new file mode 100644 index 0000000..d7ba41b --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i1024_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 49.60790970898233, + "request_tps": 1.6126460572378174, + "input_tps": 1651.349562611525, + "output_tps": 412.83739065288125, + "total_tps": 2064.1869532644064, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 935.7815920375288, + "ttft_p95_ms": 987.9027201328427, + "ttft_p99_ms": 988.8454519142397, + "tpot_p50_ms": 35.24840932302907, + "tpot_p95_ms": 35.34214854765428, + "tpot_p99_ms": 35.575809728372484, + "e2e_p50_ms": 9909.394073998556, + "e2e_p95_ms": 9992.725722829346, + "e2e_p99_ms": 9993.638517349027, + "itl_p50_ms": 33.14175654668361, + "itl_p95_ms": 51.57231901539489, + "itl_p99_ms": 59.3762440350838, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i1024_o4096_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i1024_o4096_a1.json new file mode 100644 index 0000000..6576ad0 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i1024_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 751.3376580928452, + "request_tps": 0.1064767606658605, + "input_tps": 109.03220292184115, + "output_tps": 436.1288116873646, + "total_tps": 545.1610146092057, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 1016.063955030404, + "ttft_p95_ms": 1555.9732719440944, + "ttft_p99_ms": 1557.605766323395, + "tpot_p50_ms": 36.36836581573491, + "tpot_p95_ms": 36.73897300841474, + "tpot_p99_ms": 36.74271235461387, + "e2e_p50_ms": 149854.72987149842, + "e2e_p95_ms": 151522.32534443028, + "e2e_p99_ms": 151525.3242665343, + "itl_p50_ms": 34.41172291059047, + "itl_p95_ms": 56.844778708182275, + "itl_p99_ms": 61.78853524383158, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i1024_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i1024_o512_a1.json new file mode 100644 index 0000000..b8dfecd --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i1024_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 94.76783368107863, + "request_tps": 0.8441682888861142, + "input_tps": 864.4283278193809, + "output_tps": 432.21416390969046, + "total_tps": 1296.6424917290713, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 947.5187105126679, + "ttft_p95_ms": 1040.429971262347, + "ttft_p99_ms": 1041.2525061913766, + "tpot_p50_ms": 35.295242745540456, + "tpot_p95_ms": 35.35704588877044, + "tpot_p99_ms": 35.459834945952224, + "e2e_p50_ms": 18922.931105131283, + "e2e_p95_ms": 19052.50374833122, + "e2e_p99_ms": 19055.464482815005, + "itl_p50_ms": 33.14921457786113, + "itl_p95_ms": 51.78932250710204, + "itl_p99_ms": 59.60216530133038, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i16384_o1024_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i16384_o1024_a1.json new file mode 100644 index 0000000..44d582c --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i16384_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 423.16662176605314, + "request_tps": 0.1890508274639578, + "input_tps": 3097.4087571694845, + "output_tps": 193.58804732309278, + "total_tps": 3290.996804492577, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 24554.481841041707, + "ttft_p95_ms": 47767.74098525056, + "ttft_p99_ms": 49284.376522910316, + "tpot_p50_ms": 58.8259498166778, + "tpot_p95_ms": 78.8021439746958, + "tpot_p99_ms": 81.38845075320822, + "e2e_p50_ms": 87506.93495490123, + "e2e_p95_ms": 88091.98981691152, + "e2e_p99_ms": 88095.21754639223, + "itl_p50_ms": 35.09103413671255, + "itl_p95_ms": 58.61164742382244, + "itl_p99_ms": 64.02320196386427, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i16384_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i16384_o128_a1.json new file mode 100644 index 0000000..545af89 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i16384_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 256.7601454681717, + "request_tps": 0.31157483516037693, + "input_tps": 5104.842099267616, + "output_tps": 39.88157890052825, + "total_tps": 5144.723678168144, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 25187.957516987808, + "ttft_p95_ms": 47894.87515146611, + "ttft_p99_ms": 49477.59126278339, + "tpot_p50_ms": 198.17720644458336, + "tpot_p95_ms": 368.67252661503744, + "tpot_p99_ms": 391.46561684295915, + "e2e_p50_ms": 54196.20531098917, + "e2e_p95_ms": 54292.84199364483, + "e2e_p99_ms": 54294.34734713286, + "itl_p50_ms": 35.08518589660525, + "itl_p95_ms": 58.37633586488664, + "itl_p99_ms": 64.08695708261803, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i16384_o2048_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i16384_o2048_a1.json new file mode 100644 index 0000000..9561b67 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i16384_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 614.5522862081416, + "request_tps": 0.13017606767620898, + "input_tps": 2132.804692807008, + "output_tps": 266.600586600876, + "total_tps": 2399.4052794078843, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 24607.956594438292, + "ttft_p95_ms": 47697.99113486661, + "ttft_p99_ms": 49292.4090649304, + "tpot_p50_ms": 48.10298325354711, + "tpot_p95_ms": 58.110758199371965, + "tpot_p99_ms": 59.46979306338756, + "e2e_p50_ms": 125611.4937895909, + "e2e_p95_ms": 126450.29024835676, + "e2e_p99_ms": 126453.36987355025, + "itl_p50_ms": 35.09154496714473, + "itl_p95_ms": 58.45766137354075, + "itl_p99_ms": 64.08793116686866, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i16384_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i16384_o256_a1.json new file mode 100644 index 0000000..89b1789 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i16384_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 278.9136179748457, + "request_tps": 0.2868271566690406, + "input_tps": 4699.376134865561, + "output_tps": 73.42775210727439, + "total_tps": 4772.803886972835, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 24543.6039295746, + "ttft_p95_ms": 47646.79518254706, + "ttft_p99_ms": 49249.08797713695, + "tpot_p50_ms": 123.02242240768072, + "tpot_p95_ms": 202.41332247026955, + "tpot_p99_ms": 213.29141370800562, + "e2e_p50_ms": 58673.51271153893, + "e2e_p95_ms": 59028.76297895564, + "e2e_p99_ms": 59031.15820220439, + "itl_p50_ms": 35.10860598180443, + "itl_p95_ms": 59.34609065297991, + "itl_p99_ms": 63.67699921829625, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i16384_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i16384_o512_a1.json new file mode 100644 index 0000000..8058862 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i16384_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 326.68043872504495, + "request_tps": 0.2448876348771317, + "input_tps": 4012.2390098269257, + "output_tps": 125.38246905709143, + "total_tps": 4137.621478884017, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 24534.058870980516, + "ttft_p95_ms": 47585.56273026624, + "ttft_p99_ms": 49259.400079401676, + "tpot_p50_ms": 80.12083880425615, + "tpot_p95_ms": 119.69057329970916, + "tpot_p99_ms": 125.067007240863, + "e2e_p50_ms": 68319.87011944875, + "e2e_p95_ms": 68537.59080437012, + "e2e_p99_ms": 68539.8418288771, + "itl_p50_ms": 35.064953030087054, + "itl_p95_ms": 58.41207688208669, + "itl_p99_ms": 63.36872420506552, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i4096_o1024_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i4096_o1024_a1.json new file mode 100644 index 0000000..e68b46c --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i4096_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 219.81785867293365, + "request_tps": 0.363937673139796, + "input_tps": 1490.6887091806045, + "output_tps": 372.67217729515113, + "total_tps": 1863.3608864757555, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 3086.4613954909146, + "ttft_p95_ms": 11646.606297139078, + "ttft_p99_ms": 11760.603103360627, + "tpot_p50_ms": 37.052429024333115, + "tpot_p95_ms": 44.42799982865398, + "tpot_p99_ms": 45.902305760968, + "e2e_p50_ms": 43634.44979744963, + "e2e_p95_ms": 49578.70842582779, + "e2e_p99_ms": 49580.79158608103, + "itl_p50_ms": 34.58560153376311, + "itl_p95_ms": 56.85673162806779, + "itl_p99_ms": 61.24365669908002, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i4096_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i4096_o128_a1.json new file mode 100644 index 0000000..3f3026f --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i4096_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 81.84067250299267, + "request_tps": 0.9775090740740753, + "input_tps": 4003.8771674074123, + "output_tps": 125.12116148148164, + "total_tps": 4128.998328888893, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 7960.9383844072, + "ttft_p95_ms": 11711.344682960771, + "ttft_p99_ms": 11753.230824810453, + "tpot_p50_ms": 66.47893063738708, + "tpot_p95_ms": 110.99464661427078, + "tpot_p99_ms": 112.9407165098686, + "e2e_p50_ms": 16375.667929998599, + "e2e_p95_ms": 16515.728015080094, + "e2e_p99_ms": 16517.64271946624, + "itl_p50_ms": 34.67924799770117, + "itl_p95_ms": 57.22848706645891, + "itl_p99_ms": 62.60195288574323, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i4096_o2048_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i4096_o2048_a1.json new file mode 100644 index 0000000..f39b007 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i4096_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 407.2602603479754, + "request_tps": 0.19643458443906506, + "input_tps": 804.5960578624105, + "output_tps": 402.29802893120524, + "total_tps": 1206.8940867936158, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 3240.20006461069, + "ttft_p95_ms": 11741.664571419824, + "ttft_p99_ms": 11840.223265609238, + "tpot_p50_ms": 36.97641234489819, + "tpot_p95_ms": 40.52301196099482, + "tpot_p99_ms": 41.266952687252264, + "e2e_p50_ms": 80944.1796640167, + "e2e_p95_ms": 87096.1231123074, + "e2e_p99_ms": 87098.59982100548, + "itl_p50_ms": 34.466654527932405, + "itl_p95_ms": 56.807562068570405, + "itl_p99_ms": 61.39076543040574, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i4096_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i4096_o256_a1.json new file mode 100644 index 0000000..22b6830 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i4096_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 78.74828192405403, + "request_tps": 1.0158951794929716, + "input_tps": 4161.106655203212, + "output_tps": 260.0691659502007, + "total_tps": 4421.1758211534125, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 3091.937036952004, + "ttft_p95_ms": 11642.126731504686, + "ttft_p99_ms": 11714.910287000239, + "tpot_p50_ms": 37.18031867425524, + "tpot_p95_ms": 68.26280832984575, + "tpot_p99_ms": 73.90357631190186, + "e2e_p50_ms": 15660.35121458117, + "e2e_p95_ms": 21110.056800360326, + "e2e_p99_ms": 21111.51674753055, + "itl_p50_ms": 34.62257899809629, + "itl_p95_ms": 57.151046535000205, + "itl_p99_ms": 62.4327381211333, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i4096_o4096_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i4096_o4096_a1.json new file mode 100644 index 0000000..8a493ce --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i4096_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 781.9880918110721, + "request_tps": 0.10230334814270799, + "input_tps": 419.0345139925319, + "output_tps": 419.0345139925319, + "total_tps": 838.0690279850638, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 3097.252030042, + "ttft_p95_ms": 11633.983958070166, + "ttft_p99_ms": 11746.600034988951, + "tpot_p50_ms": 36.86781218838315, + "tpot_p95_ms": 38.49309932329363, + "tpot_p99_ms": 38.90570051973158, + "e2e_p50_ms": 155683.48012503702, + "e2e_p95_ms": 162603.52390279295, + "e2e_p99_ms": 162604.7951752739, + "itl_p50_ms": 34.45248049683869, + "itl_p95_ms": 56.79855167400092, + "itl_p99_ms": 61.650155324023224, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i4096_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i4096_o512_a1.json new file mode 100644 index 0000000..60e3c51 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i4096_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 126.69021714199334, + "request_tps": 0.63146154300404, + "input_tps": 2586.466480144548, + "output_tps": 323.3083100180685, + "total_tps": 2909.7747901626167, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 3073.0733450036496, + "ttft_p95_ms": 11019.555248494726, + "ttft_p99_ms": 11065.65095537575, + "tpot_p50_ms": 38.32130294193836, + "tpot_p95_ms": 53.03848213837399, + "tpot_p99_ms": 58.19536544753461, + "e2e_p50_ms": 25832.93817844242, + "e2e_p95_ms": 30626.584287162405, + "e2e_p99_ms": 30627.329029503744, + "itl_p50_ms": 34.66650005429983, + "itl_p95_ms": 56.83219095226377, + "itl_p99_ms": 62.253236598335214, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i65536_o1024_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i65536_o1024_a1.json new file mode 100644 index 0000000..1b02781 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i65536_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 1330.2602253910154, + "request_tps": 0.06013860932847547, + "input_tps": 3941.2439009509685, + "output_tps": 61.58193595235888, + "total_tps": 4002.8258369033274, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 118513.18043493666, + "ttft_p95_ms": 230933.78256635042, + "ttft_p99_ms": 232188.58536266722, + "tpot_p50_ms": 137.7084055620283, + "tpot_p95_ms": 237.46356306838157, + "tpot_p99_ms": 250.69030466432596, + "e2e_p50_ms": 271416.1142479861, + "e2e_p95_ms": 272710.0276091136, + "e2e_p99_ms": 272711.17356425617, + "itl_p50_ms": 36.84138902463019, + "itl_p95_ms": 61.22844358906149, + "itl_p99_ms": 67.15500121004875, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i65536_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i65536_o128_a1.json new file mode 100644 index 0000000..67f2585 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i65536_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 1164.4257685828488, + "request_tps": 0.06870339197093095, + "input_tps": 4502.545496206931, + "output_tps": 8.794034172279162, + "total_tps": 4511.33953037921, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 120625.26360049378, + "ttft_p95_ms": 230919.13024774985, + "ttft_p99_ms": 233213.97572936953, + "tpot_p50_ms": 847.999912252151, + "tpot_p95_ms": 1666.0538613161907, + "tpot_p99_ms": 1748.680016982211, + "e2e_p50_ms": 236925.28252699412, + "e2e_p95_ms": 241650.54931361228, + "e2e_p99_ms": 241654.8121647397, + "itl_p50_ms": 36.83545487001538, + "itl_p95_ms": 60.686023067682996, + "itl_p99_ms": 88.91421182081103, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i65536_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i65536_o256_a1.json new file mode 100644 index 0000000..2392e90 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i65536_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 1179.4044030888472, + "request_tps": 0.06783084732470125, + "input_tps": 4445.362410271621, + "output_tps": 17.36469691512352, + "total_tps": 4462.727107186744, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 118520.80365340225, + "ttft_p95_ms": 219668.39720717393, + "ttft_p99_ms": 232060.81151018385, + "tpot_p50_ms": 456.4013590040964, + "tpot_p95_ms": 834.1273255841166, + "tpot_p99_ms": 887.1042354636113, + "e2e_p50_ms": 241645.95780789386, + "e2e_p95_ms": 242439.91744236555, + "e2e_p99_ms": 242443.70790595654, + "itl_p50_ms": 36.84396401513368, + "itl_p95_ms": 61.602629092521966, + "itl_p99_ms": 67.18409784138203, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i65536_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i65536_o512_a1.json new file mode 100644 index 0000000..5573e96 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c16_i65536_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 80, + "failed": 0, + "duration_s": 1229.7762659930158, + "request_tps": 0.0650524832949202, + "input_tps": 4263.27954521589, + "output_tps": 33.30687144699914, + "total_tps": 4296.586416662889, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 117602.89832751732, + "ttft_p95_ms": 220174.25819333873, + "ttft_p99_ms": 232176.57839990687, + "tpot_p50_ms": 250.0348996674941, + "tpot_p95_ms": 436.4520301115598, + "tpot_p99_ms": 462.66546465371283, + "e2e_p50_ms": 251397.16331893578, + "e2e_p95_ms": 252849.9990378623, + "e2e_p99_ms": 252851.82005017297, + "itl_p50_ms": 36.86228208243847, + "itl_p95_ms": 60.973369050771, + "itl_p99_ms": 67.14509757934138, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i16384_o1024_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i16384_o1024_a1.json new file mode 100644 index 0000000..3dacd46 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i16384_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 150.84414941398427, + "request_tps": 0.03314679435314225, + "input_tps": 543.0770786818827, + "output_tps": 33.942317417617666, + "total_tps": 577.0193960995003, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 3190.0415259879082, + "ttft_p95_ms": 3315.6291431281716, + "ttft_p99_ms": 3326.89755034633, + "tpot_p50_ms": 26.902348835946206, + "tpot_p95_ms": 26.929367462743006, + "tpot_p99_ms": 26.932126218243663, + "e2e_p50_ms": 30694.867955055088, + "e2e_p95_ms": 30857.89103670977, + "e2e_p99_ms": 30877.166467448696, + "itl_p50_ms": 24.886570870876312, + "itl_p95_ms": 45.76608098577708, + "itl_p99_ms": 50.4940598970279, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i16384_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i16384_o128_a1.json new file mode 100644 index 0000000..1e36f82 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i16384_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 30.202961774077266, + "request_tps": 0.16554667841520834, + "input_tps": 2712.3167791547735, + "output_tps": 21.189974837146668, + "total_tps": 2733.5067539919205, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 3189.695225097239, + "ttft_p95_ms": 3287.1102603152394, + "ttft_p99_ms": 3298.7277361750603, + "tpot_p50_ms": 27.00854029953773, + "tpot_p95_ms": 27.13199242340826, + "tpot_p99_ms": 27.15187601517798, + "e2e_p50_ms": 6534.987838938832, + "e2e_p95_ms": 6732.873298088089, + "e2e_p99_ms": 6747.015990102664, + "itl_p50_ms": 24.80782102793455, + "itl_p95_ms": 45.76855925843107, + "itl_p99_ms": 49.268019353039556, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i16384_o2048_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i16384_o2048_a1.json new file mode 100644 index 0000000..28f72c8 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i16384_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 288.6096170139499, + "request_tps": 0.017324440022933558, + "input_tps": 283.8436253357434, + "output_tps": 35.480453166967926, + "total_tps": 319.32407850271136, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 3130.5205351673067, + "ttft_p95_ms": 3326.790696242824, + "ttft_p99_ms": 3348.429289693013, + "tpot_p50_ms": 26.910882942859878, + "tpot_p95_ms": 26.939971389955993, + "tpot_p99_ms": 26.940729667464, + "e2e_p50_ms": 58267.43282494135, + "e2e_p95_ms": 58423.72469645925, + "e2e_p99_ms": 58437.077996963635, + "itl_p50_ms": 24.81332910247147, + "itl_p95_ms": 48.45747402869165, + "itl_p99_ms": 51.16548259742558, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i16384_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i16384_o256_a1.json new file mode 100644 index 0000000..da643dc --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i16384_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 47.27609982597642, + "request_tps": 0.10576168546908539, + "input_tps": 1732.799454725495, + "output_tps": 27.07499148008586, + "total_tps": 1759.8744462055809, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 3095.8918230608106, + "ttft_p95_ms": 3272.758443793282, + "ttft_p99_ms": 3304.284312753007, + "tpot_p50_ms": 27.063766945445657, + "tpot_p95_ms": 27.16058336417465, + "tpot_p99_ms": 27.17001092872199, + "e2e_p50_ms": 9997.152394149452, + "e2e_p95_ms": 10198.707201657817, + "e2e_p99_ms": 10232.637099577114, + "itl_p50_ms": 24.803407955914736, + "itl_p95_ms": 48.14038993790746, + "itl_p99_ms": 50.25792038068175, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i16384_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i16384_o512_a1.json new file mode 100644 index 0000000..a070944 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i16384_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 81.74180358601734, + "request_tps": 0.06116821235463043, + "input_tps": 1002.179991218265, + "output_tps": 31.31812472557078, + "total_tps": 1033.4981159438357, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 3152.1379488985986, + "ttft_p95_ms": 3280.934850871563, + "ttft_p99_ms": 3303.358238041401, + "tpot_p50_ms": 26.90626537553062, + "tpot_p95_ms": 26.921899768354372, + "tpot_p99_ms": 26.923089013626313, + "e2e_p50_ms": 16891.912296880037, + "e2e_p95_ms": 17034.689639275894, + "e2e_p99_ms": 17060.389525359496, + "itl_p50_ms": 24.79128399863839, + "itl_p95_ms": 48.06917109526691, + "itl_p99_ms": 50.64208318479359, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i4096_o1024_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i4096_o1024_a1.json new file mode 100644 index 0000000..abc5b72 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i4096_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 137.4567740750499, + "request_tps": 0.0363750716081119, + "input_tps": 148.99229330682635, + "output_tps": 37.24807332670659, + "total_tps": 186.24036663353294, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 250.88916602544487, + "ttft_p95_ms": 253.6237696185708, + "ttft_p99_ms": 253.63623090088367, + "tpot_p50_ms": 26.61167698223162, + "tpot_p95_ms": 26.682842084191165, + "tpot_p99_ms": 26.683082640242063, + "e2e_p50_ms": 27477.38489904441, + "e2e_p95_ms": 27547.797128511593, + "e2e_p99_ms": 27548.662202609703, + "itl_p50_ms": 24.606115883216262, + "itl_p95_ms": 47.627381072379656, + "itl_p99_ms": 50.5407242104411, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i4096_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i4096_o128_a1.json new file mode 100644 index 0000000..a108222 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i4096_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 18.347577747888863, + "request_tps": 0.2725155368574643, + "input_tps": 1116.2236389681739, + "output_tps": 34.88198871775543, + "total_tps": 1151.1056276859294, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 250.82844705320895, + "ttft_p95_ms": 256.7141311708838, + "ttft_p99_ms": 257.63806560076773, + "tpot_p50_ms": 26.909995739766227, + "tpot_p95_ms": 27.076458471907873, + "tpot_p99_ms": 27.091512166432977, + "e2e_p50_ms": 3668.39790600352, + "e2e_p95_ms": 3695.424357103184, + "e2e_p99_ms": 3698.260110737756, + "itl_p50_ms": 24.74083099514246, + "itl_p95_ms": 47.6031189551577, + "itl_p99_ms": 49.66991530265658, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i4096_o2048_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i4096_o2048_a1.json new file mode 100644 index 0000000..5d2888c --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i4096_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 273.5867548820097, + "request_tps": 0.018275738539156837, + "input_tps": 74.8574250563864, + "output_tps": 37.4287125281932, + "total_tps": 112.28613758457962, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 261.7046369705349, + "ttft_p95_ms": 349.8987853061408, + "ttft_p99_ms": 362.0478419493884, + "tpot_p50_ms": 26.592822536383906, + "tpot_p95_ms": 26.61543222684295, + "tpot_p99_ms": 26.61759176356131, + "e2e_p50_ms": 54736.32405605167, + "e2e_p95_ms": 54748.984178341925, + "e2e_p99_ms": 54749.01293158531, + "itl_p50_ms": 24.60544602945447, + "itl_p95_ms": 47.4006032804027, + "itl_p99_ms": 49.956067339517176, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i4096_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i4096_o256_a1.json new file mode 100644 index 0000000..f3c2188 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i4096_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 35.27845326298848, + "request_tps": 0.1417295696817192, + "input_tps": 580.5243174163219, + "output_tps": 36.28276983852012, + "total_tps": 616.807087254842, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 268.0197979789227, + "ttft_p95_ms": 291.9198756571859, + "ttft_p99_ms": 292.9051029961556, + "tpot_p50_ms": 26.539895318302456, + "tpot_p95_ms": 26.827862831928275, + "tpot_p99_ms": 26.862892086781997, + "e2e_p50_ms": 7060.824715998024, + "e2e_p95_ms": 7112.919608317316, + "e2e_p99_ms": 7118.816237747669, + "itl_p50_ms": 24.650061037391424, + "itl_p95_ms": 44.851708575151854, + "itl_p99_ms": 49.219665322452784, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i4096_o4096_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i4096_o4096_a1.json new file mode 100644 index 0000000..d651564 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i4096_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 548.1068854129408, + "request_tps": 0.009122308318081111, + "input_tps": 37.36497487086023, + "output_tps": 37.36497487086023, + "total_tps": 74.72994974172046, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 270.29989496804774, + "ttft_p95_ms": 280.77137977816164, + "ttft_p99_ms": 282.4734351504594, + "tpot_p50_ms": 26.703159420515647, + "tpot_p95_ms": 26.721241031938273, + "tpot_p99_ms": 26.721443862750927, + "e2e_p50_ms": 109619.73772197962, + "e2e_p95_ms": 109701.13868489861, + "e2e_p99_ms": 109702.0101480931, + "itl_p50_ms": 24.71006289124489, + "itl_p95_ms": 45.58566838968543, + "itl_p99_ms": 49.87936850637197, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i4096_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i4096_o512_a1.json new file mode 100644 index 0000000..e63cb01 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i4096_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 69.34047695202753, + "request_tps": 0.07210795511918958, + "input_tps": 295.3541841682005, + "output_tps": 36.91927302102506, + "total_tps": 332.27345718922555, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 248.92402300611138, + "ttft_p95_ms": 266.0276424139738, + "ttft_p99_ms": 268.3548469096422, + "tpot_p50_ms": 26.661601453862808, + "tpot_p95_ms": 26.700651652677212, + "tpot_p99_ms": 26.707691252833037, + "e2e_p50_ms": 13878.469962859526, + "e2e_p95_ms": 13891.861443966627, + "e2e_p99_ms": 13893.72499614954, + "itl_p50_ms": 24.624896002933383, + "itl_p95_ms": 45.65126097295369, + "itl_p99_ms": 49.571911790408194, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i65536_o1024_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i65536_o1024_a1.json new file mode 100644 index 0000000..82f7676 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i65536_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 201.46440248796716, + "request_tps": 0.024818280243323058, + "input_tps": 1626.49081402642, + "output_tps": 25.41391896916281, + "total_tps": 1651.9047329955827, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 14628.885317826644, + "ttft_p95_ms": 14868.979735858738, + "ttft_p99_ms": 14913.150340802968, + "tpot_p50_ms": 27.836642022547792, + "tpot_p95_ms": 27.844008100505842, + "tpot_p99_ms": 27.84499085022142, + "e2e_p50_ms": 43054.00717793964, + "e2e_p95_ms": 43330.51605997607, + "e2e_p99_ms": 43373.7068128027, + "itl_p50_ms": 25.475302943959832, + "itl_p95_ms": 50.770317669957876, + "itl_p99_ms": 54.30494689382613, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i65536_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i65536_o128_a1.json new file mode 100644 index 0000000..ad69822 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i65536_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 76.5485741731245, + "request_tps": 0.0653180030328436, + "input_tps": 4280.680646760438, + "output_tps": 8.360704388203981, + "total_tps": 4289.041351148642, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 14596.343518933281, + "ttft_p95_ms": 14810.734046529979, + "ttft_p99_ms": 14839.107076432556, + "tpot_p50_ms": 27.468161299956712, + "tpot_p95_ms": 27.539072871046862, + "tpot_p99_ms": 27.552164999592144, + "e2e_p50_ms": 18084.800004027784, + "e2e_p95_ms": 18259.232654888183, + "e2e_p99_ms": 18286.724362093955, + "itl_p50_ms": 25.447878055274487, + "itl_p95_ms": 46.60883932374417, + "itl_p99_ms": 51.75249738618731, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i65536_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i65536_o256_a1.json new file mode 100644 index 0000000..1757a72 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i65536_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 94.3989737580996, + "request_tps": 0.052966677506607864, + "input_tps": 3471.224177073053, + "output_tps": 13.559469441691613, + "total_tps": 3484.7836465147443, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 14599.749388871714, + "ttft_p95_ms": 14722.214041789994, + "ttft_p99_ms": 14730.883328346536, + "tpot_p50_ms": 27.881732368476545, + "tpot_p95_ms": 27.903486282619486, + "tpot_p99_ms": 27.905507876263826, + "e2e_p50_ms": 21712.56085904315, + "e2e_p95_ms": 21806.284560635686, + "e2e_p99_ms": 21809.799600169063, + "itl_p50_ms": 25.44583403505385, + "itl_p95_ms": 49.89882283844054, + "itl_p99_ms": 52.76037988718599, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i65536_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i65536_o512_a1.json new file mode 100644 index 0000000..f09bea2 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c1_i65536_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 5, + "failed": 0, + "duration_s": 130.4824561579153, + "request_tps": 0.03831932772593422, + "input_tps": 2511.295461846825, + "output_tps": 19.619495795678322, + "total_tps": 2530.9149576425034, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 14583.937218878418, + "ttft_p95_ms": 14795.338988443837, + "ttft_p99_ms": 14826.081829732284, + "tpot_p50_ms": 27.982182919691134, + "tpot_p95_ms": 28.00536987685221, + "tpot_p99_ms": 28.00938046322279, + "e2e_p50_ms": 28872.23194888793, + "e2e_p95_ms": 29090.035974793136, + "e2e_p99_ms": 29125.418794117868, + "itl_p50_ms": 25.67037590779364, + "itl_p95_ms": 50.41482762899249, + "itl_p99_ms": 53.483981639146805, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c32_i1024_o1024_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c32_i1024_o1024_a1.json new file mode 100644 index 0000000..671234e --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c32_i1024_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 235.65979528496973, + "request_tps": 0.6789448314954245, + "input_tps": 695.2395074513147, + "output_tps": 695.2395074513147, + "total_tps": 1390.4790149026294, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 2981.6380555275828, + "ttft_p95_ms": 5691.65476138005, + "ttft_p99_ms": 5695.868596371729, + "tpot_p50_ms": 42.80786131479516, + "tpot_p95_ms": 45.161864508836935, + "tpot_p99_ms": 45.66859811666182, + "e2e_p50_ms": 47101.06354195159, + "e2e_p95_ms": 49224.63555838913, + "e2e_p99_ms": 49227.704358948395, + "itl_p50_ms": 40.25210742838681, + "itl_p95_ms": 59.54401966882868, + "itl_p99_ms": 67.25824875058606, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c32_i1024_o2048_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c32_i1024_o2048_a1.json new file mode 100644 index 0000000..2f2208a --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c32_i1024_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 455.1479200378526, + "request_tps": 0.35153406827981004, + "input_tps": 359.9708859185255, + "output_tps": 719.941771837051, + "total_tps": 1079.9126577555764, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 2997.489857953042, + "ttft_p95_ms": 4408.118210406974, + "ttft_p99_ms": 4409.418380276766, + "tpot_p50_ms": 43.03901399925722, + "tpot_p95_ms": 43.72752447672071, + "tpot_p99_ms": 43.95574388355626, + "e2e_p50_ms": 91594.98853096738, + "e2e_p95_ms": 92511.16100725485, + "e2e_p99_ms": 92513.411904776, + "itl_p50_ms": 40.91528104618192, + "itl_p95_ms": 61.7140918620862, + "itl_p99_ms": 67.57471586111933, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c32_i1024_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c32_i1024_o256_a1.json new file mode 100644 index 0000000..211cb67 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c32_i1024_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 72.94000257505104, + "request_tps": 2.1935836900385253, + "input_tps": 2246.22969859945, + "output_tps": 561.5574246498625, + "total_tps": 2807.7871232493126, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 2986.9613859336823, + "ttft_p95_ms": 5638.966712448744, + "ttft_p99_ms": 5717.785984254442, + "tpot_p50_ms": 43.25931125649196, + "tpot_p95_ms": 52.76475823822194, + "tpot_p99_ms": 60.14857301456559, + "e2e_p50_ms": 14733.08369051665, + "e2e_p95_ms": 16383.586864836978, + "e2e_p99_ms": 16391.066697821952, + "itl_p50_ms": 40.04870797507465, + "itl_p95_ms": 61.18088683579117, + "itl_p99_ms": 67.89261806756254, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c32_i1024_o4096_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c32_i1024_o4096_a1.json new file mode 100644 index 0000000..1a19825 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c32_i1024_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 898.5537388259545, + "request_tps": 0.17806391881364284, + "input_tps": 182.33745286517026, + "output_tps": 729.349811460681, + "total_tps": 911.6872643258514, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 3016.2806854350492, + "ttft_p95_ms": 3879.865962965414, + "ttft_p99_ms": 3881.218241781462, + "tpot_p50_ms": 43.30877847119516, + "tpot_p95_ms": 43.5328147706794, + "tpot_p99_ms": 43.585850752513146, + "e2e_p50_ms": 179637.5885854941, + "e2e_p95_ms": 181239.3637237721, + "e2e_p99_ms": 181245.79159579705, + "itl_p50_ms": 40.978090022690594, + "itl_p95_ms": 61.92717199446633, + "itl_p99_ms": 67.58996407035738, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c32_i1024_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c32_i1024_o512_a1.json new file mode 100644 index 0000000..3f1045a --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c32_i1024_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 160, + "failed": 0, + "duration_s": 116.47414470394142, + "request_tps": 1.3736954274847377, + "input_tps": 1406.6641177443714, + "output_tps": 703.3320588721857, + "total_tps": 2109.996176616557, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 1728.7794339936227, + "ttft_p95_ms": 1780.919153173454, + "ttft_p99_ms": 1782.5818243529648, + "tpot_p50_ms": 42.263241610289306, + "tpot_p95_ms": 42.444828835041704, + "tpot_p99_ms": 42.508655317581514, + "e2e_p50_ms": 23282.43359492626, + "e2e_p95_ms": 23419.968224375043, + "e2e_p99_ms": 23422.798421205953, + "itl_p50_ms": 40.06265802308917, + "itl_p95_ms": 58.03677138173953, + "itl_p99_ms": 67.34221158549191, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c48_i1024_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c48_i1024_o512_a1.json new file mode 100644 index 0000000..6ff5985 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c48_i1024_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 240, + "failed": 0, + "duration_s": 538.8399277571589, + "request_tps": 0.44540129199215867, + "input_tps": 456.0909229999705, + "output_tps": 228.04546149998524, + "total_tps": 684.1363844999557, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 2837.4407634837553, + "ttft_p95_ms": 8041.054585680831, + "ttft_p99_ms": 8290.32544139307, + "tpot_p50_ms": 202.26081837298543, + "tpot_p95_ms": 209.53028491557413, + "tpot_p99_ms": 212.22897237452446, + "e2e_p50_ms": 106162.95209643431, + "e2e_p95_ms": 111388.48089557141, + "e2e_p99_ms": 111393.2417512266, + "itl_p50_ms": 197.14024697896093, + "itl_p95_ms": 226.85785598587245, + "itl_p99_ms": 233.4809394343756, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i1024_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i1024_o128_a1.json new file mode 100644 index 0000000..8616aa5 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i1024_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 23.175778079079464, + "request_tps": 1.7259398956752865, + "input_tps": 1767.3624531714934, + "output_tps": 220.92030664643667, + "total_tps": 1988.28275981793, + "mean_input_tokens": 1024.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 586.4921135362238, + "ttft_p95_ms": 622.7808163734153, + "ttft_p99_ms": 623.125078517478, + "tpot_p50_ms": 31.85279993340373, + "tpot_p95_ms": 32.22969000018752, + "tpot_p99_ms": 32.44051498659762, + "e2e_p50_ms": 4641.481174970977, + "e2e_p95_ms": 4670.240912691224, + "e2e_p99_ms": 4670.475004990585, + "itl_p50_ms": 29.695762554183602, + "itl_p95_ms": 49.66246752301232, + "itl_p99_ms": 57.65656981850043, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i16384_o1024_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i16384_o1024_a1.json new file mode 100644 index 0000000..c45713c --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i16384_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 288.24989658407867, + "request_tps": 0.13876847996832684, + "input_tps": 2273.582775801067, + "output_tps": 142.09892348756668, + "total_tps": 2415.6816992886334, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 14033.50293962285, + "ttft_p95_ms": 24491.35324291419, + "ttft_p99_ms": 24759.231015304103, + "tpot_p50_ms": 41.92089678206183, + "tpot_p95_ms": 52.53722619232089, + "tpot_p99_ms": 52.69310522824981, + "e2e_p50_ms": 58040.07501259912, + "e2e_p95_ms": 58448.35202746326, + "e2e_p99_ms": 58450.28124490753, + "itl_p50_ms": 30.73286300059408, + "itl_p95_ms": 50.72684104088693, + "itl_p99_ms": 57.92215333785862, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i16384_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i16384_o128_a1.json new file mode 100644 index 0000000..8500c91 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i16384_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 141.1061867091339, + "request_tps": 0.283474459432832, + "input_tps": 4644.44554334752, + "output_tps": 36.2847308074025, + "total_tps": 4680.730274154922, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 14016.848506988026, + "ttft_p95_ms": 24515.615026152227, + "ttft_p99_ms": 24870.744992317632, + "tpot_p50_ms": 105.65934030661785, + "tpot_p95_ms": 191.19213953821398, + "tpot_p99_ms": 193.04246251707838, + "e2e_p50_ms": 28676.901251543313, + "e2e_p95_ms": 29117.543387238402, + "e2e_p99_ms": 29119.045873775613, + "itl_p50_ms": 30.759848072193563, + "itl_p95_ms": 50.73059933492914, + "itl_p99_ms": 57.838678041007384, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i16384_o2048_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i16384_o2048_a1.json new file mode 100644 index 0000000..e34d1e6 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i16384_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 457.1198655979242, + "request_tps": 0.08750440094673856, + "input_tps": 1433.6721051113645, + "output_tps": 179.20901313892057, + "total_tps": 1612.881118250285, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 14047.149811987765, + "ttft_p95_ms": 24552.844235277735, + "ttft_p99_ms": 24749.485357583035, + "tpot_p50_ms": 37.538202461908504, + "tpot_p95_ms": 42.66441069203499, + "tpot_p99_ms": 42.9089602763204, + "e2e_p50_ms": 91750.90344448108, + "e2e_p95_ms": 92103.39087992907, + "e2e_p99_ms": 92105.43445243035, + "itl_p50_ms": 30.728504993021488, + "itl_p95_ms": 51.14136257907376, + "itl_p99_ms": 58.35309494053946, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i16384_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i16384_o256_a1.json new file mode 100644 index 0000000..90b29f5 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i16384_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 162.15329257398844, + "request_tps": 0.24668015903376442, + "input_tps": 4041.607725609196, + "output_tps": 63.15012071264369, + "total_tps": 4104.75784632184, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 14056.083535542712, + "ttft_p95_ms": 24572.28690140182, + "ttft_p99_ms": 24781.60693206126, + "tpot_p50_ms": 69.22504512739239, + "tpot_p95_ms": 111.92492204695465, + "tpot_p99_ms": 112.1908097408584, + "e2e_p50_ms": 32945.15154813416, + "e2e_p95_ms": 33157.10190224927, + "e2e_p99_ms": 33157.680223954376, + "itl_p50_ms": 30.78671346884221, + "itl_p95_ms": 52.047876664437354, + "itl_p99_ms": 57.51584505662322, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i16384_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i16384_o512_a1.json new file mode 100644 index 0000000..14e5fa5 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i16384_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 204.11131394607946, + "request_tps": 0.19597149823143506, + "input_tps": 3210.797027023832, + "output_tps": 100.33740709449475, + "total_tps": 3311.134434118327, + "mean_input_tokens": 16384.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 14109.92513189558, + "ttft_p95_ms": 24601.094605890103, + "ttft_p99_ms": 24752.130593049806, + "tpot_p50_ms": 50.97825450885589, + "tpot_p95_ms": 72.14095964029964, + "tpot_p99_ms": 72.50778853802152, + "e2e_p50_ms": 41347.00059145689, + "e2e_p95_ms": 41603.95504550543, + "e2e_p99_ms": 41605.48491582973, + "itl_p50_ms": 30.709165963344276, + "itl_p95_ms": 51.023998111486435, + "itl_p99_ms": 57.31804177630693, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i4096_o1024_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i4096_o1024_a1.json new file mode 100644 index 0000000..b31d1ff --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i4096_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 196.3215377200395, + "request_tps": 0.2037473853584074, + "input_tps": 834.5492904280367, + "output_tps": 208.63732260700917, + "total_tps": 1043.1866130350459, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 5105.625024996698, + "ttft_p95_ms": 5958.576378412545, + "ttft_p99_ms": 6064.859193183947, + "tpot_p50_ms": 33.55810541205767, + "tpot_p95_ms": 36.34212366745689, + "tpot_p99_ms": 36.55665332609869, + "e2e_p50_ms": 39269.71339690499, + "e2e_p95_ms": 39593.57306190068, + "e2e_p99_ms": 39594.54825331457, + "itl_p50_ms": 30.5747389793396, + "itl_p95_ms": 50.719153042882674, + "itl_p99_ms": 56.99894693214446, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i4096_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i4096_o128_a1.json new file mode 100644 index 0000000..24ab7ce --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i4096_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 49.58700227085501, + "request_tps": 0.8066629997415711, + "input_tps": 3304.091646941475, + "output_tps": 103.2528639669211, + "total_tps": 3407.3445109083964, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 5025.09073051624, + "ttft_p95_ms": 5915.067831950728, + "ttft_p99_ms": 5934.279049527831, + "tpot_p50_ms": 39.28249386619542, + "tpot_p95_ms": 62.01645719083613, + "tpot_p99_ms": 62.2953052225332, + "e2e_p50_ms": 10006.435959017836, + "e2e_p95_ms": 10091.394168324769, + "e2e_p99_ms": 10092.632040372118, + "itl_p50_ms": 30.577882542274892, + "itl_p95_ms": 49.62789301062003, + "itl_p99_ms": 57.752098962664604, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i4096_o2048_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i4096_o2048_a1.json new file mode 100644 index 0000000..8c1ad96 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i4096_o2048_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 363.24991058488376, + "request_tps": 0.11011702641741698, + "input_tps": 451.03934020573996, + "output_tps": 225.51967010286998, + "total_tps": 676.5590103086099, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 2048.0, + "ttft_p50_ms": 5055.334709468298, + "ttft_p95_ms": 5906.876079260837, + "ttft_p99_ms": 6023.611045922153, + "tpot_p50_ms": 33.129859284787166, + "tpot_p95_ms": 34.53019272802082, + "tpot_p99_ms": 34.57053666370664, + "e2e_p50_ms": 72728.55528548826, + "e2e_p95_ms": 73087.71414748626, + "e2e_p99_ms": 73088.92423222307, + "itl_p50_ms": 30.50815698225051, + "itl_p95_ms": 50.1406642026268, + "itl_p99_ms": 56.95488589582965, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i4096_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i4096_o256_a1.json new file mode 100644 index 0000000..0b74966 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i4096_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 70.42232585605234, + "request_tps": 0.5680016885804441, + "input_tps": 2326.534916425499, + "output_tps": 145.4084322765937, + "total_tps": 2471.943348702093, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 5035.245958482847, + "ttft_p95_ms": 5984.106912557036, + "ttft_p99_ms": 6042.94591952581, + "tpot_p50_ms": 35.78011738455982, + "tpot_p95_ms": 46.93394760007733, + "tpot_p99_ms": 47.50900067536927, + "e2e_p50_ms": 14144.402876030654, + "e2e_p95_ms": 14412.95803807443, + "e2e_p99_ms": 14413.140350417234, + "itl_p50_ms": 30.597370583564043, + "itl_p95_ms": 50.928640994243324, + "itl_p99_ms": 56.54409846989438, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i4096_o4096_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i4096_o4096_a1.json new file mode 100644 index 0000000..9935df5 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i4096_o4096_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 697.7969978589099, + "request_tps": 0.057323261812152054, + "input_tps": 234.79608038257481, + "output_tps": 234.79608038257481, + "total_tps": 469.59216076514963, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 4096.0, + "ttft_p50_ms": 5059.23117196653, + "ttft_p95_ms": 6020.107874658424, + "ttft_p99_ms": 6025.99322974449, + "tpot_p50_ms": 32.921226551871115, + "tpot_p95_ms": 33.65806957458343, + "tpot_p99_ms": 33.68422295676342, + "e2e_p50_ms": 139956.67370257434, + "e2e_p95_ms": 140090.03641876625, + "e2e_p99_ms": 140090.8339079097, + "itl_p50_ms": 30.493864906020463, + "itl_p95_ms": 50.366084929555655, + "itl_p99_ms": 57.08304369123653, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i4096_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i4096_o512_a1.json new file mode 100644 index 0000000..a0afdcb --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i4096_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 112.55219594296068, + "request_tps": 0.3553906671023215, + "input_tps": 1455.6801724511088, + "output_tps": 181.9600215563886, + "total_tps": 1637.6401940074975, + "mean_input_tokens": 4096.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 5065.312996157445, + "ttft_p95_ms": 6016.085800586734, + "ttft_p99_ms": 6074.749361765571, + "tpot_p50_ms": 34.33472787484767, + "tpot_p95_ms": 40.08658976554269, + "tpot_p99_ms": 40.18997862394681, + "e2e_p50_ms": 22591.4841449121, + "e2e_p95_ms": 22795.752924506087, + "e2e_p99_ms": 22796.309465037193, + "itl_p50_ms": 30.596577562391758, + "itl_p95_ms": 50.246595905628055, + "itl_p99_ms": 56.772089656442404, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i65536_o1024_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i65536_o1024_a1.json new file mode 100644 index 0000000..e7bb451 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i65536_o1024_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 739.3616181469988, + "request_tps": 0.054100725569510504, + "input_tps": 3545.5451509234404, + "output_tps": 55.399142983178756, + "total_tps": 3600.9442939066194, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 1024.0, + "ttft_p50_ms": 60294.08791498281, + "ttft_p95_ms": 116841.35833900655, + "ttft_p99_ms": 116967.29428508785, + "tpot_p50_ms": 82.37472311334132, + "tpot_p95_ms": 131.37778040708542, + "tpot_p99_ms": 132.5520263926535, + "e2e_p50_ms": 150254.07721009105, + "e2e_p95_ms": 151760.6172351283, + "e2e_p99_ms": 151763.06750814896, + "itl_p50_ms": 31.72238008119166, + "itl_p95_ms": 53.61441783607006, + "itl_p99_ms": 61.79662807378918, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i65536_o128_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i65536_o128_a1.json new file mode 100644 index 0000000..933f9d6 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i65536_o128_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 587.0072881928645, + "request_tps": 0.06814225445674156, + "input_tps": 4465.770788077015, + "output_tps": 8.72220857046292, + "total_tps": 4474.492996647478, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 128.0, + "ttft_p50_ms": 60264.91823955439, + "ttft_p95_ms": 116942.56678273669, + "ttft_p99_ms": 116968.76564948354, + "tpot_p50_ms": 419.5593314018012, + "tpot_p95_ms": 818.3646683727837, + "tpot_p99_ms": 829.4111081236225, + "e2e_p50_ms": 119654.99441663269, + "e2e_p95_ms": 121336.96261633886, + "e2e_p99_ms": 121339.05124852201, + "itl_p50_ms": 31.66217007674277, + "itl_p95_ms": 56.3370821531862, + "itl_p99_ms": 61.347583890892565, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i65536_o256_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i65536_o256_a1.json new file mode 100644 index 0000000..44cf69e --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i65536_o256_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 608.7442546100356, + "request_tps": 0.06570903905388674, + "input_tps": 4306.307583435521, + "output_tps": 16.821513997795005, + "total_tps": 4323.129097433316, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 256.0, + "ttft_p50_ms": 60408.123059547506, + "ttft_p95_ms": 116816.32042041747, + "ttft_p99_ms": 117158.49739520112, + "tpot_p50_ms": 225.85036875497477, + "tpot_p95_ms": 424.8813510864662, + "tpot_p99_ms": 430.2625958223681, + "e2e_p50_ms": 123890.77430358157, + "e2e_p95_ms": 125879.0796181187, + "e2e_p99_ms": 125880.81930729328, + "itl_p50_ms": 31.708508962765336, + "itl_p95_ms": 55.004988797009, + "itl_p99_ms": 60.499521191231906, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i65536_o512_a1.json b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i65536_o512_a1.json new file mode 100644 index 0000000..cf82421 --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/metrics/sglang_c8_i65536_o512_a1.json @@ -0,0 +1,25 @@ +{ + "status": "COMPLETED", + "completed": 40, + "failed": 0, + "duration_s": 652.6434917969164, + "request_tps": 0.061289203834498415, + "input_tps": 4016.649262497688, + "output_tps": 31.38007236326319, + "total_tps": 4048.029334860951, + "mean_input_tokens": 65536.0, + "mean_output_tokens": 512.0, + "ttft_p50_ms": 60481.61956341937, + "ttft_p95_ms": 116843.18738668226, + "ttft_p99_ms": 117324.80349385645, + "tpot_p50_ms": 131.188325970543, + "tpot_p95_ms": 229.34224289243244, + "tpot_p99_ms": 231.58556326105642, + "e2e_p50_ms": 132822.63546145987, + "e2e_p95_ms": 134675.67310340237, + "e2e_p99_ms": 134677.4176377803, + "itl_p50_ms": 31.73411381430924, + "itl_p95_ms": 53.03707402199507, + "itl_p99_ms": 61.14733610767871, + "validation_errors": [] +} diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/server_cmd.txt b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/server_cmd.txt new file mode 100644 index 0000000..cff8c7e --- /dev/null +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/adaptive_results/adaptive_20260716-074243/tp8_dp1/server_cmd.txt @@ -0,0 +1 @@ +--host 0.0.0.0 --port 30014 --model-path /models --attention-backend nsa --nsa-prefill klxdsa --nsa-decode klxdsa --trust-remote-code --disable-custom-all-reduce --page-size 64 --mem-fraction-static 0.8 --tensor-parallel-size 8 --disable-shared-experts-fusion --quantization w8a8_int8 --kv-cache-dtype float16 --disable-piecewise-cuda-graph --cuda-graph-max-bs 32 --watchdog-timeout 3000000 --tool-call-parser deepseekv4 --reasoning-parser deepseek-v4 --constrained-json-disable-any-whitespace --enable-metrics --enable-request-time-stats-logging --context-length 140000 diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/run_adaptive_concurrency.sh b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/run_adaptive_concurrency.sh index a18ad88..16b6056 100755 --- a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/run_adaptive_concurrency.sh +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/run_adaptive_concurrency.sh @@ -7,15 +7,15 @@ SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" EXPERIMENT_NAME="$(basename "$SCRIPT_DIR")" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/lib.sh" +source "${SCRIPT_DIR}/../../../scripts/common/lib.sh" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/platform.sh" +source "${SCRIPT_DIR}/../../../scripts/common/platform.sh" # shellcheck source=/dev/null source "${SCRIPT_DIR}/config.env" # shellcheck source=/dev/null source "${SCRIPT_DIR}/adaptive_config.env" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/adaptive_bench_lib.sh" +source "${SCRIPT_DIR}/../../../scripts/common/adaptive_bench_lib.sh" ENGINE="sglang" ENGINE_PORT="$SGLANG_PORT" diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/run_adaptive_concurrency_add16.sh b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/run_adaptive_concurrency_add16.sh index dc82e78..91176a6 100755 --- a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/run_adaptive_concurrency_add16.sh +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/run_adaptive_concurrency_add16.sh @@ -7,15 +7,15 @@ SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" EXPERIMENT_NAME="$(basename "$SCRIPT_DIR")" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/lib.sh" +source "${SCRIPT_DIR}/../../../scripts/common/lib.sh" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/platform.sh" +source "${SCRIPT_DIR}/../../../scripts/common/platform.sh" # shellcheck source=/dev/null source "${SCRIPT_DIR}/config.env" # shellcheck source=/dev/null source "${SCRIPT_DIR}/adaptive_config.env" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/adaptive_bench_lib.sh" +source "${SCRIPT_DIR}/../../../scripts/common/adaptive_bench_lib.sh" ENGINE="sglang" ENGINE_PORT="$SGLANG_PORT" diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/run_bench.sh b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/run_bench.sh index bdbe1e5..bb643c8 100755 --- a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/run_bench.sh +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/run_bench.sh @@ -6,9 +6,9 @@ SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" EXPERIMENT_NAME="$(basename "$SCRIPT_DIR")" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/lib.sh" +source "${SCRIPT_DIR}/../../../scripts/common/lib.sh" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/platform.sh" +source "${SCRIPT_DIR}/../../../scripts/common/platform.sh" # shellcheck source=/dev/null source "${SCRIPT_DIR}/config.env" @@ -519,7 +519,7 @@ PY # Parse results. log "parsing ${config_label} results" - "$PYTHON" "${SCRIPT_DIR}/../../../../scripts/common/parse_backend.py" "$result_root" --backend sglang \ + "$PYTHON" "${SCRIPT_DIR}/../../../scripts/common/parse_backend.py" "$result_root" --backend sglang \ >> "${phase_log_dir}/parse.log" 2>&1 || { log "WARNING: parser failed for ${config_label}; see ${phase_log_dir}/parse.log" } diff --git a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/start_sglang_docker.sh b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/start_sglang_docker.sh index c2b7a13..f19ebb4 100755 --- a/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/start_sglang_docker.sh +++ b/experiments/p800/dsv4_p800_sglang_tp_dp_matrix/start_sglang_docker.sh @@ -15,9 +15,9 @@ fi SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/lib.sh" +source "${SCRIPT_DIR}/../../../scripts/common/lib.sh" # shellcheck source=/dev/null -source "${SCRIPT_DIR}/../../../../scripts/common/platform.sh" +source "${SCRIPT_DIR}/../../../scripts/common/platform.sh" # shellcheck source=/dev/null source "${SCRIPT_DIR}/config.env"