A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 12, 19:45 UTC
[CI Bugfix] Fix Kernels DeepGEMM Test (H100) (#29106)
Signed-off-by: mgoin <mgoin64@gmail.com>
Michael Goin committed
Nov 20, 2025 at 19:42 UTC
986ab5db6325fb4a5d937084ca7921a95641504a
1 file changed
+4
-5
.buildkite/test-pipeline.yaml
+4
-5
@@ -554,7 +554,6 @@ steps:
554
timeout_in_minutes: 45
555
gpu: h100
556
num_gpus: 1
557
- optional: true
557
source_file_dependencies:
558
- tools/install_deepgemm.sh
559
- vllm/utils/deep_gemm.py
@@ -565,10 +564,10 @@ steps:
564
- tests/kernels/moe/test_batched_deepgemm.py
565
- tests/kernels/attention/test_deepgemm_attention.py
566
commands:
568
- - pytest -v -s tests/kernels/quantization/test_block_fp8.py -k deep_gemm
569
- - pytest -v -s tests/kernels/moe/test_deepgemm.py
570
- - pytest -v -s tests/kernels/moe/test_batched_deepgemm.py
571
- - pytest -v -s tests/kernels/attention/test_deepgemm_attention.py
567
+ - pytest -v -s kernels/quantization/test_block_fp8.py -k deep_gemm
568
+ - pytest -v -s kernels/moe/test_deepgemm.py
569
+ - pytest -v -s kernels/moe/test_batched_deepgemm.py
570
+ - pytest -v -s kernels/attention/test_deepgemm_attention.py
571
572
- label: Model Executor Test # 23min
573
timeout_in_minutes: 35