docker run --rm --network host -v /data/hf_models/Kimi-K3:/data/hf_models/Kimi-K3:ro -v /data/hzy/sskj/experiments/pro6000/kimi3_pro6000_pd_pp8_standard/results/kimi3-pd-decode512-20260827-1325:/data/hzy/sskj/experiments/pro6000/kimi3_pro6000_pd_pp8_standard/results/kimi3-pd-decode512-20260827-1325 --entrypoint python3 local/sglang:kimi-k3-sm120-flashinfer-mxfp4-pd1 -m sglang.benchmark.serving --backend sglang --host 174.1.60.5 --port 31000 --tokenizer /data/hf_models/Kimi-K3 --dataset-name random-ids --random-input-len 16384 --random-output-len 512 --random-range-ratio 1.0 --num-prompts 1 --max-concurrency 1 --request-rate 10000 --warmup-requests 0 --output-file /data/hzy/sskj/experiments/pro6000/kimi3_pro6000_pd_pp8_standard/results/kimi3-pd-decode512-20260827-1325/bench/decode_16k_512_c1.jsonl --output-details --disable-tqdm 
