Remove Quantization test parallelism (#49693)
Signed-off-by: khluu <khluu000@gmail.com> Co-authored-by: OpenAI Codex <noreply@openai.com>
Kevin H. Luu committed
Jul 24, 2026 at 01:14 UTC
5c5434e2d89fe7008ec7a6046faa65a69ba568ca
1 file changed
+2
-4
.buildkite/test_areas/quantization.yaml
+2
-4
@@ -24,8 +24,7 @@ steps:
24
- uv pip install --system conch-triton-kernels
25
# The SM90-only checkpoint currently contains a removed weight_chan_scale
26
# parameter. It was not exercised by the previous L4 job.
27
- - VLLM_TEST_FORCE_LOAD_FORMAT=auto pytest -v -s quantization/ --ignore quantization/test_blackwell_moe.py -k 'not test_compressed_tensors_w4a8_fp8' --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT
28
- parallelism: 8
27
+ - VLLM_TEST_FORCE_LOAD_FORMAT=auto pytest -v -s quantization/ --ignore quantization/test_blackwell_moe.py -k 'not test_compressed_tensors_w4a8_fp8'
28
29
- label: Quantized Fusions
30
device: h200_35gb
@@ -68,5 +67,4 @@ steps:
67
- vllm/model_executor/layers/quantization
68
- tests/models/quantization
69
commands:
71
- - pytest -v -s models/quantization --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT
72
- parallelism: 3
70
+ - pytest -v -s models/quantization