From da9134bc89364465f68aa679564435936d157a14 Mon Sep 17 00:00:00 2001 From: Andreas Karatzas Date: Sat, 17 Oct 2026 21:49:27 -0500 Subject: [PATCH] [ROCm][CI] Give the MoE parallel-launch watchdog room for AITER JIT builds (#62161) Signed-off-by: Andreas Karatzas Co-authored-by: Andreas Karatzas Co-authored-by: Claude --- tests/kernels/moe/modular_kernel_tools/parallel_utils.py | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/tests/kernels/moe/modular_kernel_tools/parallel_utils.py b/tests/kernels/moe/modular_kernel_tools/parallel_utils.py index 235d1b81a4..b6cafe4a41 100644 --- a/tests/kernels/moe/modular_kernel_tools/parallel_utils.py +++ b/tests/kernels/moe/modular_kernel_tools/parallel_utils.py @@ -18,13 +18,16 @@ from vllm.distributed import ( init_distributed_environment, initialize_model_parallel, ) +from vllm.platforms import current_platform from vllm.utils.network_utils import get_open_port ## Parallel Processes Utils P = ParamSpec("P") -LAUNCH_TIMEOUT_S = 900 +# On ROCm the first AITER CK MoE launch JIT-compiles two stage modules that the +# AITER wheel does not prebuild, ~10 min each (ROCm/aiter#5766). +LAUNCH_TIMEOUT_S = 1800 if current_platform.is_rocm() else 900 @dataclasses.dataclass