@tduncan / Nvidia-Optimization-Stack-VLLM Mirror · read-only / commits / 5478c4b41f
Fork 0

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, 15:45 UTC

[perf bench] set timeout to debug hanging (#8516)

Simon Mo committed Sep 16, 2024 at 14:30 UTC 5478c4b41f60995b92b9997306b2e0702055341f
2 files changed +4 -3
.buildkite/nightly-benchmarks/benchmark-pipeline.yaml
+1 -2
@@ -8,8 +8,7 @@ steps:
8 containers:
9 - image: badouralix/curl-jq
10 command:
11 - - sh
12 - - .buildkite/nightly-benchmarks/scripts/wait-for-image.sh
11 + - sh .buildkite/nightly-benchmarks/scripts/wait-for-image.sh
12 - wait
13 - label: "A100"
14 agents:
.buildkite/nightly-benchmarks/scripts/wait-for-image.sh
+3 -1
@@ -2,9 +2,11 @@
2 TOKEN=$(curl -s -L "https://public.ecr.aws/token?service=public.ecr.aws&scope=repository:q9t5s3a7/vllm-ci-test-repo:pull" | jq -r .token)
3 URL="https://public.ecr.aws/v2/q9t5s3a7/vllm-ci-test-repo/manifests/$BUILDKITE_COMMIT"
4
5 +TIMEOUT_SECONDS=10
6 +
7 retries=0
8 while [ $retries -lt 1000 ]; do
7 - if [ $(curl -s -L -H "Authorization: Bearer $TOKEN" -o /dev/null -w "%{http_code}" $URL) -eq 200 ]; then
9 + if [ $(curl -s --max-time $TIMEOUT_SECONDS -L -H "Authorization: Bearer $TOKEN" -o /dev/null -w "%{http_code}" $URL) -eq 200 ]; then
10 exit 0
11 fi
12