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 16, 05:45 UTC
[benchmark] Make H100 benchmark optional (#10908)
Kevin H. Luu committed
Dec 4, 2024 at 17:02 UTC
7883c2bbe7d0ab47160d205822f7b188a5a2771b
1 file changed
+5
.buildkite/nightly-benchmarks/benchmark-pipeline.yaml
+5
@@ -65,10 +65,15 @@ steps:
65
- VLLM_USAGE_SOURCE
66
- HF_TOKEN
67
68
+ - block: "Run H100 Benchmark"
69
+ key: block-h100
70
+ depends_on: ~
71
+
72
- label: "H100"
73
# skip: "use this flag to conditionally skip the benchmark step, useful for PR testing"
74
agents:
75
queue: H100
76
+ depends_on: block-h100
77
plugins:
78
- docker#v5.12.0:
79
image: public.ecr.aws/q9t5s3a7/vllm-ci-postmerge-repo:$BUILDKITE_COMMIT