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, 01:15 UTC
Change targets for AMD build in the "CI" pipeline (#34918)
Signed-off-by: Alexei V. Ivanov <alexei.ivanov@amd.com>
Alexei-V-Ivanov-AMD committed
Feb 19, 2026 at 15:26 UTC
304319c4edcc1a50317c22715ad6c0111459025d
1 file changed
+1
-1
.buildkite/hardware_tests/amd.yaml
+1
-1
@@ -10,7 +10,7 @@ steps:
10
docker build
11
--build-arg max_jobs=16
12
--build-arg REMOTE_VLLM=1
13
- --build-arg ARG_PYTORCH_ROCM_ARCH='gfx90a;gfx942'
13
+ --build-arg ARG_PYTORCH_ROCM_ARCH='gfx942;gfx950'
14
--build-arg VLLM_BRANCH=$BUILDKITE_COMMIT
15
--tag "rocm/vllm-ci:${BUILDKITE_COMMIT}"
16
-f docker/Dockerfile.rocm