From daeffd147b218ac3beb3142d2229f096b8c1ca53 Mon Sep 17 00:00:00 2001
From: Zhiyi Hong <2497491955@qq.com>
Date: Tue, 18 Aug 2026 12:36:47 +0800
Subject: [PATCH] [Fix] Use built-in random IDs for Kimi Prefill matrix

---
 .../README.md                                               | 6 ++++--
 .../config.env                                              | 3 +--
 .../run_moe_backend_prefill.sh                              | 6 +++---
 3 files changed, 8 insertions(+), 7 deletions(-)

diff --git a/experiments/pro6000/kimi3_pro6000_sglang_tp32ep32_moe_backend_prefill/README.md b/experiments/pro6000/kimi3_pro6000_sglang_tp32ep32_moe_backend_prefill/README.md
index cd5e012..2b3a0aa 100644
--- a/experiments/pro6000/kimi3_pro6000_sglang_tp32ep32_moe_backend_prefill/README.md
+++ b/experiments/pro6000/kimi3_pro6000_sglang_tp32ep32_moe_backend_prefill/README.md
@@ -17,8 +17,10 @@ MXFP4 MoE on NVIDIA SM120. It does not modify the existing baseline experiment.
 | Repeats | 3 |
 
 The eight configurations are identical except for MoE runner and chunk size.
-Radix cache is disabled. `OSL=1` makes TTFT, E2E latency and input TPS the
-primary metrics; TPOT is not meaningful for this matrix.
+The client uses SGLang `random-ids`, which produces the exact synthetic token
+length without depending on a ShareGPT file or text truncation. Radix cache is
+disabled. `OSL=1` makes TTFT, E2E latency and input TPS the primary metrics;
+TPOT is not meaningful for this matrix.
 
 ## Unique entrypoint
 
diff --git a/experiments/pro6000/kimi3_pro6000_sglang_tp32ep32_moe_backend_prefill/config.env b/experiments/pro6000/kimi3_pro6000_sglang_tp32ep32_moe_backend_prefill/config.env
index 7f067cf..f1e8b18 100644
--- a/experiments/pro6000/kimi3_pro6000_sglang_tp32ep32_moe_backend_prefill/config.env
+++ b/experiments/pro6000/kimi3_pro6000_sglang_tp32ep32_moe_backend_prefill/config.env
@@ -17,7 +17,7 @@ FLASHINFER_TARGET="${FLASHINFER_TARGET:-/tmp/hzy-fi-situ-phase3-v3}"
 FLASHINFER_WHEEL="${FLASHINFER_WHEEL:-/data/hzy/artifacts/flashinfer-kimi-sm120-situ-phase3-v3/flashinfer_python-0.6.18-py3-none-any.whl}"
 SGLANG_SOURCE="${SGLANG_SOURCE:-/data/hzy/src/sglang-kimi-sm120/python/sglang}"
 
-DATASET_PATH="${DATASET_PATH:-${ROOT_DIR}/dataset/ShareGPT_V3_unfiltered_cleaned_split.json}"
+BENCH_DATASET_NAME="${BENCH_DATASET_NAME:-random-ids}"
 BENCH_MODULE="${BENCH_MODULE:-sglang.benchmark.serving}"
 RESULT_BASE="${RESULT_BASE:-${SCRIPT_DIR}/results}"
 
@@ -31,4 +31,3 @@ HEALTH_WAIT_S="${HEALTH_WAIT_S:-2400}"
 BACKENDS=(marlin flashinfer_mxfp4)
 CHUNKED_PREFILL_SIZES=(8192 16384)
 CONCURRENCIES=(8 16)
-
diff --git a/experiments/pro6000/kimi3_pro6000_sglang_tp32ep32_moe_backend_prefill/run_moe_backend_prefill.sh b/experiments/pro6000/kimi3_pro6000_sglang_tp32ep32_moe_backend_prefill/run_moe_backend_prefill.sh
index e22792f..ed6cb7d 100755
--- a/experiments/pro6000/kimi3_pro6000_sglang_tp32ep32_moe_backend_prefill/run_moe_backend_prefill.sh
+++ b/experiments/pro6000/kimi3_pro6000_sglang_tp32ep32_moe_backend_prefill/run_moe_backend_prefill.sh
@@ -78,7 +78,7 @@ sudo_host() {
 
 check_inputs() {
   local path
-  for path in "$MODEL_PATH" "$DATASET_PATH" "$FLASHINFER_TARGET/flashinfer" \
+  for path in "$MODEL_PATH" "$FLASHINFER_TARGET/flashinfer" \
     "$FLASHINFER_WHEEL" "$SGLANG_SOURCE"; do
     [[ -e "$path" ]] || { echo "ERROR: required path missing: $path" >&2; exit 2; }
   done
@@ -254,16 +254,16 @@ run_one_bench() {
   local stem="${backend}_chunk${chunk}_c${concurrency}_r${repeat}"
   local output_file="${RESULT_ROOT}/raw/${stem}.jsonl"
   local bench_log="${RESULT_ROOT}/bench/${stem}.log"
+  local -a dataset_args=(--dataset-name "$BENCH_DATASET_NAME")
   rm -f "$output_file"
   log "bench backend=${backend} chunk=${chunk} C=${concurrency} repeat=${repeat}/${REPEATS}"
   sudo_host "$HEAD_HOST" docker run --rm --network host \
     -v "${MODEL_PATH}:${MODEL_PATH}:ro" \
-    -v "${DATASET_PATH}:${DATASET_PATH}:ro" \
     -v "${RESULT_ROOT}:${RESULT_ROOT}" \
     -e PYTHONUNBUFFERED=1 --entrypoint python3 "$CANDIDATE_IMAGE" \
     -m "$BENCH_MODULE" \
     --backend sglang --host "$HEAD_HOST" --port "$PORT" \
-    --dataset-name random --dataset-path "$DATASET_PATH" \
+    "${dataset_args[@]}" \
     --random-input-len "$INPUT_LEN" --random-output-len "$OUTPUT_LEN" \
     --random-range-ratio 1.0 --num-prompts "$NUM_PROMPTS" \
     --max-concurrency "$concurrency" --request-rate 10000 \