From a332b84578cdc0706e040f6a765954c8a289904f Mon Sep 17 00:00:00 2001
From: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Date: Wed, 1 Oct 2025 10:03:44 +0100
Subject: [PATCH] [CI] Only capture a single CUDA graph size in CI by default
 (#25951)

Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
---
 tests/conftest.py | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/tests/conftest.py b/tests/conftest.py
index a50985a465e6..e3e412f90af5 100644
--- a/tests/conftest.py
+++ b/tests/conftest.py
@@ -731,6 +731,9 @@ class VllmRunner:
         init_ctx = (nullcontext() if default_torch_num_threads is None else
                     set_default_torch_num_threads(default_torch_num_threads))
 
+        if not kwargs.get("compilation_config", None):
+            kwargs["compilation_config"] = {"cudagraph_capture_sizes": [8]}
+
         with init_ctx:
             self.llm = LLM(
                 model=model_name,