From 72c3e638cd73209c93fc7a04c864b89c467f6ae9 Mon Sep 17 00:00:00 2001 From: 1am9trash <1am9trash@gmail.com> Date: Wed, 28 Oct 2026 03:13:23 +0000 Subject: [PATCH 1/4] Update image version and args --- .../srt-slurm-recipes/dsv4/sglang/mi355x-fp4-mtp/agentic.yaml | 1 + inferencex-e2e/configs/amd-master.yaml | 2 +- 2 files changed, 2 insertions(+), 1 deletion(-) diff --git a/inferencex-e2e/benchmarks/single_node/srt-slurm-recipes/dsv4/sglang/mi355x-fp4-mtp/agentic.yaml b/inferencex-e2e/benchmarks/single_node/srt-slurm-recipes/dsv4/sglang/mi355x-fp4-mtp/agentic.yaml index 826eaa1f0f..884971a138 100644 --- a/inferencex-e2e/benchmarks/single_node/srt-slurm-recipes/dsv4/sglang/mi355x-fp4-mtp/agentic.yaml +++ b/inferencex-e2e/benchmarks/single_node/srt-slurm-recipes/dsv4/sglang/mi355x-fp4-mtp/agentic.yaml @@ -88,6 +88,7 @@ override_tp8_c1: max-running-requests: 2 cuda-graph-max-bs-decode: 2 cuda-graph-backend-prefill: breakable + enable-amd-monokernel: true benchmark: env: CONC: '1' diff --git a/inferencex-e2e/configs/amd-master.yaml b/inferencex-e2e/configs/amd-master.yaml index 99774cd09a..95be884fdc 100644 --- a/inferencex-e2e/configs/amd-master.yaml +++ b/inferencex-e2e/configs/amd-master.yaml @@ -1537,7 +1537,7 @@ glm5.2-fp4-mi355x-atom-agentic-mtp-mono: - { tp: 4, kv-offloading: none, conc-list: [1, 2], spec-decoding: mtp, srt-recipe: agentic.yaml } dsv4-fp4-mi355x-sglang-agentic-mtp: - image: lmsysorg/sglang-rocm:v0.5.22-rocm720-mi35x-20261019 + image: lmsysorg/sglang-rocm:v0.5.22-rocm10-mi35x-20261028 model: deepseek-ai/DeepSeek-V4-Pro-0813 model-prefix: dsv4 runner: cluster:mi355x-amds From 6757fb8d9f8dc2964929af138ae652898fbebfe8 Mon Sep 17 00:00:00 2001 From: 1am9trash <1am9trash@gmail.com> Date: Wed, 28 Oct 2026 03:13:23 +0000 Subject: [PATCH 2/4] Update changelog --- inferencex-e2e/perf-changelog.yaml | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/inferencex-e2e/perf-changelog.yaml b/inferencex-e2e/perf-changelog.yaml index 518e7d6b8d..082fab4f84 100644 --- a/inferencex-e2e/perf-changelog.yaml +++ b/inferencex-e2e/perf-changelog.yaml @@ -11471,3 +11471,12 @@ - "Engines set minimax_m3_msa_decode_backend=cutlass. In nightly-5555a2d8 the CUTLASS path engages only at 16 or more concurrent FP8-KV decode requests, so it can take effect only at C25/C30; the other points run Triton MSA decode." - "The shipped BF16 EAGLE3-GQA draft (3 tokens) uses FlashInfer attention with local argmax reduction; AgentX throughput applies the committed golden AL." pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/4281 + +- config-keys: + - dsv4-fp4-mi355x-sglang-agentic-mtp + scenario-type: + - agentic-coding + description: + - "Bump to lmsysorg/sglang-rocm:v0.5.22-rocm10-mi35x-20261028, which pins sgl-project/sgl-amd-mono@db7adb7 (sgl-project/sglang#46902)." + - "Enable --enable-amd-monokernel at concurrency 1. Each DSpark verify there is one request's 7 tokens, which the monokernels run as one persistent launch per layer; steps with more rows run the standard forward." + pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/XXX From 46ff7609b0ea9857f917f75b27b5d782ee6d4a27 Mon Sep 17 00:00:00 2001 From: 1am9trash <1am9trash@gmail.com> Date: Wed, 28 Oct 2026 09:05:14 +0000 Subject: [PATCH 3/4] Update image --- inferencex-e2e/configs/amd-master.yaml | 2 +- inferencex-e2e/perf-changelog.yaml | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/inferencex-e2e/configs/amd-master.yaml b/inferencex-e2e/configs/amd-master.yaml index ad909b3bf1..fcb4c56fdf 100644 --- a/inferencex-e2e/configs/amd-master.yaml +++ b/inferencex-e2e/configs/amd-master.yaml @@ -1537,7 +1537,7 @@ glm5.2-fp4-mi355x-atom-agentic-mtp-mono: - { tp: 4, kv-offloading: none, conc-list: [1, 2], spec-decoding: mtp, srt-recipe: agentic.yaml } dsv4-fp4-mi355x-sglang-agentic-mtp: - image: lmsysorg/sglang-rocm:v0.5.22-rocm10-mi35x-20261028 + image: lmsysorg/sglang-rocm:v0.5.22-rocm720-mi35x-20261028 model: deepseek-ai/DeepSeek-V4-Pro-0813 model-prefix: dsv4 runner: cluster:mi355x-amds diff --git a/inferencex-e2e/perf-changelog.yaml b/inferencex-e2e/perf-changelog.yaml index 0fa33c3126..8477204cec 100644 --- a/inferencex-e2e/perf-changelog.yaml +++ b/inferencex-e2e/perf-changelog.yaml @@ -11495,6 +11495,6 @@ scenario-type: - agentic-coding description: - - "Bump to lmsysorg/sglang-rocm:v0.5.22-rocm10-mi35x-20261028, which pins sgl-project/sgl-amd-mono@db7adb7 (sgl-project/sglang#46902)." + - "Bump to lmsysorg/sglang-rocm:v0.5.22-rocm720-mi35x-20261028, which pins sgl-project/sgl-amd-mono@90c2355 (sgl-project/sglang#46951)." - "Enable --enable-amd-monokernel at concurrency 1. Each DSpark verify there is one request's 7 tokens, which the monokernels run as one persistent launch per layer; steps with more rows run the standard forward." pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/4596 From 9ab9d0a58ccbcbb19e7bc82fbd0e920e26625914 Mon Sep 17 00:00:00 2001 From: Thomas Wang <1am9trash@gmail.com> Date: Wed, 28 Oct 2026 18:22:04 +0800 Subject: [PATCH 4/4] Refine description in changelog --- inferencex-e2e/perf-changelog.yaml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/inferencex-e2e/perf-changelog.yaml b/inferencex-e2e/perf-changelog.yaml index 8477204cec..e6d3f6059b 100644 --- a/inferencex-e2e/perf-changelog.yaml +++ b/inferencex-e2e/perf-changelog.yaml @@ -11495,6 +11495,6 @@ scenario-type: - agentic-coding description: - - "Bump to lmsysorg/sglang-rocm:v0.5.22-rocm720-mi35x-20261028, which pins sgl-project/sgl-amd-mono@90c2355 (sgl-project/sglang#46951)." + - "Bump to lmsysorg/sglang-rocm:v0.5.22-rocm720-mi35x-20261028." - "Enable --enable-amd-monokernel at concurrency 1. Each DSpark verify there is one request's 7 tokens, which the monokernels run as one persistent launch per layer; steps with more rows run the standard forward." pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/4596