Add server CUDA graph warmup CI step for cold H200 nodes (#19201)

2026-02-24 16:07:01 -08:00
parent c193a52fa2
commit 750ecf4a45
4 changed files with 739 additions and 7 deletions
--- a/test/registered/8-gpu-models/test_ring_2_5_1t.py
+++ b/test/registered/8-gpu-models/test_ring_2_5_1t.py
@@ -20,7 +20,6 @@ class TestRing2_5_1T(unittest.TestCase):

    def test_ring_2_5_1t(self):
        base_args = [
-            "--tp=8",
            "--trust-remote-code",
            "--model-loader-extra-config",
            '{"enable_multithread_load": true, "num_threads": 64}',