Skip to content
This repository was archived by the owner on Oct 11, 2024. It is now read-only.

Commit 1ccd388

Browse files
khluuRobert Shaw
authored andcommitted
[ci] Limit num gpus if specified for A100 (vllm-project#5694)
Signed-off-by: kevin <kevin@anyscale.com>
1 parent 0e8e31e commit 1ccd388

File tree

2 files changed

+2
-1
lines changed

2 files changed

+2
-1
lines changed

.buildkite/test-pipeline.yaml

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -184,6 +184,7 @@ steps:
184184

185185
- label: Distributed Tests (A100)
186186
gpu: a100
187+
num_gpus: 4
187188
commands:
188189
# NOTE: don't test llama model here, it seems hf implementation is buggy
189190
# see https://github.com/vllm-project/vllm/pull/5689 for details

.buildkite/test-template-aws.j2

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -75,7 +75,7 @@ steps:
7575
- "'cd {{ (step.working_dir or default_working_dir) | safe }} && {{ step.command or (step.commands | join(' && ')) | safe }}'"
7676
resources:
7777
limits:
78-
nvidia.com/gpu: 8
78+
nvidia.com/gpu: {{ step.num_gpus or 1 }}
7979
volumeMounts:
8080
- name: devshm
8181
mountPath: /dev/shm

0 commit comments

Comments
 (0)