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 8, 12:30 UTC
[Benchmarks] Fix attention benchmark smoke test (#34269)
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
Matthew Bonanni committed
Feb 10, 2026 at 16:04 UTC
4293c00b84b968ed25f80dfd2af3bb34d1eeeef6
1 file changed
+2
-1
.buildkite/test_areas/benchmarks.yaml
+2
-1
@@ -22,9 +22,10 @@ steps:
22
device: b200
23
num_gpus: 2
24
optional: true
25
+ working_dir: "/vllm-workspace/"
26
timeout_in_minutes: 10
27
source_file_dependencies:
28
- benchmarks/attention_benchmarks/
29
- vllm/v1/attention/
30
commands:
30
- - python benchmarks/attention_benchmarks/benchmark.py --backends flash flashinfer --batch-specs "8q1s1k" --repeats 1 --warmup-iters 1
31
+ - python3 benchmarks/attention_benchmarks/benchmark.py --backends flash flashinfer --batch-specs "8q1s1k" --repeats 1 --warmup-iters 1