From 0fad59514d0106c78c3770e89e4a2506eecd5665 Mon Sep 17 00:00:00 2001 From: Klaud-Cold Date: Thu, 10 Sep 2026 06:36:36 +0000 Subject: [PATCH 1/2] config: bump qwen3.5-fp4-b200-sglang SGLang image to v0.5.19-cu130 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Update the qwen3.5-fp4-b200-sglang master image from lmsysorg/sglang:v0.5.14-cu130 to lmsysorg/sglang:v0.5.19-cu130 (digest sha256:d6e7288627be8b02be88e4bba38e73f6d50e2826869f753c13a4c4385ab3eda9). Model, TP4/EP1 and TP2/EP1 topologies, 8k/1k workload and the launch script are unchanged. 将 qwen3.5-fp4-b200-sglang 的主配置镜像从 lmsysorg/sglang:v0.5.14-cu130 更新至 lmsysorg/sglang:v0.5.19-cu130（摘要 sha256:d6e7288627be8b02be88e4bba38e73f6d50e2826869f753c13a4c4385ab3eda9）。 模型、TP4/EP1 与 TP2/EP1 拓扑、8k/1k 工作负载以及启动脚本均保持不变。 Co-Authored-By: Claude Fable 5.1 --- configs/nvidia-master.yaml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/configs/nvidia-master.yaml b/configs/nvidia-master.yaml index 11fa1c1ce7..2f20b6ab4d 100644 --- a/configs/nvidia-master.yaml +++ b/configs/nvidia-master.yaml @@ -1234,7 +1234,7 @@ qwen3.5-fp8-b200-sglang-agentic-mtp: - { tp: 4, ep: 1, spec-decoding: mtp, kv-offloading: dram, kv-offload-backend: { name: hicache }, conc-list: [16, 18, 20, 22, 24, 28, 32] } qwen3.5-fp4-b200-sglang: - image: lmsysorg/sglang:v0.5.14-cu130 + image: lmsysorg/sglang:v0.5.19-cu130 model: nvidia/Qwen3.5-397B-A17B-NVFP4-V2 model-prefix: qwen3.5 runner: cluster:b200-nscale From baed6629770e3235ed43b23decaa5300e5f67930 Mon Sep 17 00:00:00 2001 From: Klaud-Cold Date: Thu, 10 Sep 2026 07:47:45 +0000 Subject: [PATCH 2/2] changelog: record qwen3.5-fp4-b200-sglang image bump to v0.5.19-cu130 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Append the perf-changelog.yaml entry for the qwen3.5-fp4-b200-sglang SGLang image update from lmsysorg/sglang:v0.5.14-cu130 to lmsysorg/sglang:v0.5.19-cu130. 为 qwen3.5-fp4-b200-sglang 的 SGLang 镜像更新（lmsysorg/sglang:v0.5.14-cu130 → lmsysorg/sglang:v0.5.19-cu130）追加 perf-changelog.yaml 条目。 Co-Authored-By: Claude Fable 5.1 --- perf-changelog.yaml | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/perf-changelog.yaml b/perf-changelog.yaml index e9fab1e0d8..f94f2333a3 100644 --- a/perf-changelog.yaml +++ b/perf-changelog.yaml @@ -7058,3 +7058,10 @@ description: - "Remove the unstable concurrency-1152 2P DEP8 prefill plus DEP12 decode point after repeated benchmark failures." pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2623 + +- config-keys: + - qwen3.5-fp4-b200-sglang + description: + - "Update SGLang image from lmsysorg/sglang:v0.5.14-cu130 to lmsysorg/sglang:v0.5.19-cu130 (digest sha256:d6e7288627be8b02be88e4bba38e73f6d50e2826869f753c13a4c4385ab3eda9)." + - "Model nvidia/Qwen3.5-397B-A17B-NVFP4-V2, TP4/EP1 and TP2/EP1 topologies, the 8k/1k workload and the launch script are unchanged." + pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2954