From 34c86c78083fbe95ad1b4958bd80bfa8106fc5e7 Mon Sep 17 00:00:00 2001 From: Chun Fang Date: Fri, 10 Apr 2026 12:42:57 +0000 Subject: [PATCH 1/5] Upgrade GLM-5 image to v0.5.10 --- .github/configs/amd-master.yaml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.github/configs/amd-master.yaml b/.github/configs/amd-master.yaml index 867a81716..db6bfa7e7 100644 --- a/.github/configs/amd-master.yaml +++ b/.github/configs/amd-master.yaml @@ -240,7 +240,7 @@ qwen3.5-fp8-mi300x-sglang: - { tp: 8, conc-start: 4, conc-end: 64 } glm5-fp8-mi355x-sglang: - image: rocm/sgl-dev:v0.5.8.post1-rocm720-mi35x-20260219 + image: lmsysorg/sglang:v0.5.10-rocm720-mi35x model: zai-org/GLM-5-FP8 model-prefix: glm5 runner: mi355x From f0bbb074b067d7d859d9e030b4a8b2b1eb52feba Mon Sep 17 00:00:00 2001 From: Chun Fang Date: Fri, 10 Apr 2026 12:48:01 +0000 Subject: [PATCH 2/5] Update perf-changelog --- perf-changelog.yaml | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/perf-changelog.yaml b/perf-changelog.yaml index 19e7988e6..0636f9154 100644 --- a/perf-changelog.yaml +++ b/perf-changelog.yaml @@ -1322,3 +1322,9 @@ description: - "Qwen3.5 fp4 support on SGL" pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1006 + +- config-keys: + - glm5-fp8-mi355x-sglang + description: + - "Upgrade GLM5 FP8 MI355X SGLang image to v0.5.10" + pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1023 From 8d32c240e0ded2682cd1fb9f287814e77c7d1dee Mon Sep 17 00:00:00 2001 From: Chun Fang Date: Mon, 13 Apr 2026 14:46:09 +0000 Subject: [PATCH 3/5] Set kv-cache-dytpe fp8-e4m3 and disable-radix-cache --- benchmarks/single_node/glm5_fp8_mi355x.sh | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/benchmarks/single_node/glm5_fp8_mi355x.sh b/benchmarks/single_node/glm5_fp8_mi355x.sh index 3d82fd856..668120e57 100755 --- a/benchmarks/single_node/glm5_fp8_mi355x.sh +++ b/benchmarks/single_node/glm5_fp8_mi355x.sh @@ -49,7 +49,9 @@ python3 -m sglang.launch_server \ --mem-fraction-static 0.85 \ --model-loader-extra-config '{"enable_multithread_load": true, "num_threads": 8}' \ --nsa-prefill-backend tilelang \ - --nsa-decode-backend tilelang $EVAL_CONTEXT_ARGS > $SERVER_LOG 2>&1 & + --nsa-decode-backend tilelang $EVAL_CONTEXT_ARGS \ + --kv-cache-dtype fp8_e4m3 \ + --disable-radix-cache> $SERVER_LOG 2>&1 & SERVER_PID=$! From 58486f74b740b9f45c69853f8442b3385a2dd757 Mon Sep 17 00:00:00 2001 From: Cameron Quilici Date: Tue, 14 Apr 2026 15:09:48 -0500 Subject: [PATCH 4/5] Update amd-master.yaml change to nightly image --- .github/configs/amd-master.yaml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.github/configs/amd-master.yaml b/.github/configs/amd-master.yaml index db6bfa7e7..5242940b7 100644 --- a/.github/configs/amd-master.yaml +++ b/.github/configs/amd-master.yaml @@ -240,7 +240,7 @@ qwen3.5-fp8-mi300x-sglang: - { tp: 8, conc-start: 4, conc-end: 64 } glm5-fp8-mi355x-sglang: - image: lmsysorg/sglang:v0.5.10-rocm720-mi35x + image: rocm/sgl-dev:v0.5.10rc0-rocm720-mi35x-20260413 model: zai-org/GLM-5-FP8 model-prefix: glm5 runner: mi355x From 7c4289f1870a3dda9ca5b7c28bbf2a9826a130ee Mon Sep 17 00:00:00 2001 From: Chun Fang Date: Wed, 15 Apr 2026 07:20:32 +0000 Subject: [PATCH 5/5] Upgrade GLM5 FP8 MI355X SGLang image to v0.5.10rc0-rocm720-mi35x-20260413 Updated the perf-changelog too --- .github/configs/amd-master.yaml | 2 +- perf-changelog.yaml | 3 ++- 2 files changed, 3 insertions(+), 2 deletions(-) diff --git a/.github/configs/amd-master.yaml b/.github/configs/amd-master.yaml index 9441e62aa..084893e11 100644 --- a/.github/configs/amd-master.yaml +++ b/.github/configs/amd-master.yaml @@ -242,7 +242,7 @@ qwen3.5-fp8-mi300x-sglang: - { tp: 8, conc-start: 4, conc-end: 64 } glm5-fp8-mi355x-sglang: - image: rocm/sgl-dev:v0.5.10rc0-rocm720-mi35x-20260413 + image: lmsysorg/sglang-rocm:v0.5.10rc0-rocm720-mi35x-20260413 model: zai-org/GLM-5-FP8 model-prefix: glm5 runner: mi355x diff --git a/perf-changelog.yaml b/perf-changelog.yaml index 5e605f9d1..73f0d86ce 100644 --- a/perf-changelog.yaml +++ b/perf-changelog.yaml @@ -1356,7 +1356,8 @@ - config-keys: - glm5-fp8-mi355x-sglang description: - - "Upgrade GLM5 FP8 MI355X SGLang image to v0.5.10" + - "Upgrade GLM5 FP8 MI355X SGLang image to v0.5.10rc0-rocm720-mi35x-20260413" + - "Set --kv-cache-dtype fp8_e4m3 and --disable-radix-cache" pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1023 - config-keys: