From ca21288080882714d7d10ee8588641225b08756c Mon Sep 17 00:00:00 2001
From: Lucas Wilkinson <LucasWilkinson@users.noreply.github.com>
Date: Fri, 16 Jan 2026 14:27:16 -0700
Subject: [PATCH] [CI] Fix OOM in Hopper Fusion E2E Tests (H100) (#32489)

Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
---
 .buildkite/test-pipeline.yaml | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/.buildkite/test-pipeline.yaml b/.buildkite/test-pipeline.yaml
index 12760def7..56bb3c23d 100644
--- a/.buildkite/test-pipeline.yaml
+++ b/.buildkite/test-pipeline.yaml
@@ -1064,7 +1064,8 @@ steps:
   - tests/compile/test_fusion_attn.py
   commands:
     - export VLLM_TEST_CLEAN_GPU_MEMORY=1
-    - pytest -v -s tests/compile/test_fusion_attn.py
+    # skip Llama-4 since it does not fit on this device
+    - pytest -v -s tests/compile/test_fusion_attn.py -k 'not Llama-4'
 
 - label: Hopper Fusion Distributed E2E Tests (2xH100)  # 70min
   timeout_in_minutes: 70