From bbee64147bf2712bf2c9dfa323de2a55b3f90a08 Mon Sep 17 00:00:00 2001 From: Oseltamivir <58582368+Oseltamivir@users.noreply.github.com> Date: Tue, 14 Jul 2026 14:05:52 +0800 Subject: [PATCH] chore(dsv4): refresh H200/B300/GB300 configs and densify B300 1k1k sweep MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Bump dsv4-fp8-h200-vllm(+mtp) image v0.21.0 -> v0.25.0 (match #2169) - Bump dsv4-fp8-h200-sglang(+mtp) deepseek-v4-hopper digest to 2026-05-13 push - Densify dsv4-fp4-b300-sglang 1k1k: DP-attn flashinfer band conc 64-512, megamoe band 2048-8192; script selects profiles by CONC range - Re-run entries for dsv4-fp4-b300-trt(+mtp) and dsv4-fp4-gb300-dynamo-trt(+mtp) 中文：刷新 H200/B300/GB300 的 DeepSeek-V4-Pro 配置：H200 vLLM 镜像升级至 v0.25.0，SGLang hopper 镜像更新至 2026-05-13 摘要；加密 B300 SGLang 1k1k 扫描点（DP-attn 64-512、megamoe 2048-8192）；追加 B300 TRT 与 GB300 dynamo-trt 重跑条目。 --- .../fixed_seq_len/dsv4_fp4_b300_sglang.sh | 4 +- configs/nvidia-master.yaml | 23 ++++++----- perf-changelog.yaml | 38 +++++++++++++++++++ 3 files changed, 53 insertions(+), 12 deletions(-) diff --git a/benchmarks/single_node/fixed_seq_len/dsv4_fp4_b300_sglang.sh b/benchmarks/single_node/fixed_seq_len/dsv4_fp4_b300_sglang.sh index b451dee0d5..553ad7279e 100755 --- a/benchmarks/single_node/fixed_seq_len/dsv4_fp4_b300_sglang.sh +++ b/benchmarks/single_node/fixed_seq_len/dsv4_fp4_b300_sglang.sh @@ -53,7 +53,7 @@ start_gpu_monitor --output "$PWD/gpu_metrics.csv" # SWA ratio: 1k inputs need more SWA cache headroom than 8k inputs; 0.5 was # tuned empirically for the 1k1k recipe, while 0.1 is the cookbook default. -if [ "$CONC" = "1" ] || [ "$CONC" = "32" ]; then +if [ "$CONC" -le 32 ]; then # TP-only, no DP attention MEM_FRACTION_STATIC=0.90 SWA_FULL_TOKENS_RATIO=$([[ "$ISL" == "1024" ]] && echo 0.5 || echo 0.1) @@ -63,7 +63,7 @@ if [ "$CONC" = "1" ] || [ "$CONC" = "32" ]; then --disable-flashinfer-autotune ) -elif [ "$CONC" = "512" ]; then +elif [ "$CONC" -ge 64 ] && [ "$CONC" -le 512 ]; then # DP attention, flashinfer_mxfp4 export SGLANG_OPT_SWA_EVICT_DROP_PAGE_MARGIN=1 MEM_FRACTION_STATIC=0.94 diff --git a/configs/nvidia-master.yaml b/configs/nvidia-master.yaml index f90c510186..20e7e589e3 100644 --- a/configs/nvidia-master.yaml +++ b/configs/nvidia-master.yaml @@ -2002,8 +2002,8 @@ dsv4-fp4-b300-sglang: multinode: false # Recipes are selected inside benchmarks/single_node/dsv4_fp4_b300_sglang.sh # by CONC: - # CONC 1|32: TP-only, flashinfer_mxfp4 - # CONC 512: DP-attn, flashinfer_mxfp4 + # CONC <=32: TP-only, flashinfer_mxfp4 + # CONC 64-512: DP-attn, flashinfer_mxfp4 # CONC 2048-8192: DP-attn, megamoe # ep is implicit in sglang: --moe-a2a-backend megamoe forces ep_size=tp_size, # while low-latency leaves ep_size at the default of 1. @@ -2014,8 +2014,12 @@ dsv4-fp4-b300-sglang: search-space: - { tp: 8, ep: 1, conc-start: 1, conc-end: 1 } - { tp: 4, ep: 1, conc-start: 32, conc-end: 32 } - - { tp: 4, ep: 1, dp-attn: true, conc-start: 512, conc-end: 512 } - - { tp: 8, ep: 8, dp-attn: true, conc-start: 8192, conc-end: 8192 } + # Densified mid band: the previous single conc-512 point left the + # ~50 tok/s/user frontier region (conc 64-256) unsampled at 1k1k. + - { tp: 4, ep: 1, dp-attn: true, conc-start: 64, conc-end: 512 } + # Megamoe band extended down to the script's existing 2048/4096 profiles + # (previously only 8192 at 1k1k). + - { tp: 8, ep: 8, dp-attn: true, conc-start: 2048, conc-end: 8192 } - isl: 8192 osl: 1024 search-space: @@ -3078,7 +3082,7 @@ dsr1-fp8-h200-sglang-mtp: # Uses the cu129 image. H200 has no FP4 path, so the FP4 indexer cache # flag is omitted. Max-model-len is pinned at 800k per the recipe. dsv4-fp8-h200-vllm: - image: vllm/vllm-openai:v0.21.0 + image: vllm/vllm-openai:v0.25.0 model: deepseek-ai/DeepSeek-V4-Pro model-prefix: dsv4 runner: h200 @@ -3098,11 +3102,10 @@ dsv4-fp8-h200-vllm: - { tp: 8, ep: 1, dp-attn: false, conc-start: 1, conc-end: 256 } - { tp: 8, ep: 8, dp-attn: true, conc-start: 1, conc-end: 256 } -# MTP variant of dsv4-fp8-h200-vllm. Uses the canonical v0.20.1 image -# (the non-MTP entry above is still on the deepseekv4-cu129 tag) and adds +# MTP variant of dsv4-fp8-h200-vllm. Mirrors the non-MTP image and adds # --speculative-config '{"method":"mtp","num_speculative_tokens":2}'. dsv4-fp8-h200-vllm-mtp: - image: vllm/vllm-openai:v0.21.0 + image: vllm/vllm-openai:v0.25.0 model: deepseek-ai/DeepSeek-V4-Pro model-prefix: dsv4 runner: h200 @@ -3143,7 +3146,7 @@ dsv4-fp8-h200-vllm-agentic: # runner pool, search space) and adds EAGLE speculative decoding via # --speculative-algorithm EAGLE with the (3,1,4) chain matching dsv4-fp4-b300-sglang-mtp. dsv4-fp8-h200-sglang: - image: lmsysorg/sglang:deepseek-v4-hopper@sha256:7f19c6dc092e47a10fac2e41f47eab78970280d06648b8e50d312a82f0ae722f + image: lmsysorg/sglang:deepseek-v4-hopper@sha256:1bf5d508ab110cc0fe1659a5f21d1be02a7f0d7ca8f58cea7e7f4e11f6ae208f model: deepseek-ai/DeepSeek-V4-Pro model-prefix: dsv4 runner: h200-dgxc @@ -3167,7 +3170,7 @@ dsv4-fp8-h200-sglang: # runner pool, search space) and adds EAGLE speculative decoding via # --speculative-algorithm EAGLE with the (3,1,4) chain matching dsv4-fp4-b300-sglang-mtp. dsv4-fp8-h200-sglang-mtp: - image: lmsysorg/sglang:deepseek-v4-hopper@sha256:7f19c6dc092e47a10fac2e41f47eab78970280d06648b8e50d312a82f0ae722f + image: lmsysorg/sglang:deepseek-v4-hopper@sha256:1bf5d508ab110cc0fe1659a5f21d1be02a7f0d7ca8f58cea7e7f4e11f6ae208f model: deepseek-ai/DeepSeek-V4-Pro model-prefix: dsv4 runner: h200-dgxc diff --git a/perf-changelog.yaml b/perf-changelog.yaml index 57d510dd15..2fefc6bd52 100644 --- a/perf-changelog.yaml +++ b/perf-changelog.yaml @@ -4750,3 +4750,41 @@ - "Image: lmsysorg/sglang:nightly-dev-cu13-20260709-074bb928" - "6 topologies across 1k/1k and 8k/1k: 1P1D TP4 STP + wide-EP (DEP4 prefill / DEP16 decode) from 1P1D up to 8P1D, recipes under benchmarks/multi_node/srt-slurm-recipes/sglang/qwen3.5/gb300-fp8/" pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2137 + +- config-keys: + - dsv4-fp8-h200-vllm + - dsv4-fp8-h200-vllm-mtp + description: + - "Bump vLLM image from v0.21.0 to v0.25.0 for DeepSeek-V4-Pro FP8 on H200, matching the B200/B300 dsv4 vLLM bump (#2169)" + - "Refresh stale H200 dsv4 submissions (last run 2026-05-21)" + pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2189 + +- config-keys: + - dsv4-fp8-h200-sglang + - dsv4-fp8-h200-sglang-mtp + description: + - "Bump the pinned lmsysorg/sglang:deepseek-v4-hopper digest from the 2026-05-02 push (7f19c6dc) to the current 2026-05-13 push (1bf5d508)" + - "Refresh stale H200 dsv4 SGLang submissions (last run 2026-05-04)" + pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2189 + +- config-keys: + - dsv4-fp4-b300-sglang + description: + - "Densify the 1k1k sweep: the previous search space sampled only conc {1, 32, 512, 8192}, leaving the ~50 tok/s/user frontier region unsampled and noisy (933 tok/s/GPU vs B200's 1359 at 50 tok/s/user)" + - "Add the DP-attn flashinfer_mxfp4 band conc 64-512 (was a single 512 point) and extend the megamoe band down to conc 2048-8192 (was a single 8192 point), reusing the existing CONC 2048/4096 launch profiles in dsv4_fp4_b300_sglang.sh" + - "dsv4_fp4_b300_sglang.sh: select launch profiles by CONC range (<=32 TP-only, 64-512 DP-attn flashinfer) instead of exact match; megamoe profiles unchanged" + pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2189 + +- config-keys: + - dsv4-fp4-b300-trt + - dsv4-fp4-b300-trt-mtp + description: + - "Re-run B300 dsv4 TRT-LLM configs (no config change, last run 2026-06-11) so the 1k1k frontier density check around 50 tok/s/user compares same-pass data across frameworks" + pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2189 + +- config-keys: + - dsv4-fp4-gb300-dynamo-trt + - dsv4-fp4-gb300-dynamo-trt-mtp + description: + - "Re-run GB300 dsv4 dynamo-trt configs (no config change): the only GB300 dsv4 1k1k submissions are from a single 2026-06-15 run" + pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2189