From 5b1b3fd14b32a983c044bedcce2515e3a91c0f69 Mon Sep 17 00:00:00 2001
From: Cam Quilici <cjquilici@gmail.com>
Date: Wed, 8 Apr 2026 14:00:47 -0500
Subject: [PATCH] Revert "[AMD] Upgrade GLM-5 SGLang mi35x image to 0.5.10
 (#1014)"

This reverts commit 6f6a5fe0796b4f15d26015426445549390b8025d.
---
 .github/configs/amd-master.yaml | 2 +-
 perf-changelog.yaml             | 9 +--------
 2 files changed, 2 insertions(+), 9 deletions(-)

diff --git a/.github/configs/amd-master.yaml b/.github/configs/amd-master.yaml
index c7688ffd1..44cc07419 100644
--- a/.github/configs/amd-master.yaml
+++ b/.github/configs/amd-master.yaml
@@ -222,7 +222,7 @@ 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:v0.5.8.post1-rocm720-mi35x-20260219
   model: zai-org/GLM-5-FP8
   model-prefix: glm5
   runner: mi355x
diff --git a/perf-changelog.yaml b/perf-changelog.yaml
index 5782e11db..6c0185ca2 100644
--- a/perf-changelog.yaml
+++ b/perf-changelog.yaml
@@ -1285,19 +1285,12 @@
     - "Runner script updated to clone NVIDIA/srt-slurm and map vLLM container image"
   pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1008
 
-- config-keys:
-    - glm5-fp8-mi355x-sglang
-  description:
-    - "Upgrade SGLang image to v0.5.10"
-    - "Resolve the issue: https://github.com/sgl-project/sglang/issues/19028"
-  pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1014
-  
 - config-keys:
     - minimaxm2.5-fp8-b200-vllm
   description:
     - "Update MiniMax-M2.5 FP8 B200 config with new search spaces"
   pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1010
-  
+
 - config-keys:
     - minimaxm2.5-fp4-b200-vllm
   description: