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, 02:00 UTC
[HPU][misc] add comments for explanation (#12034)
Signed-off-by: youkaichao <youkaichao@gmail.com>
youkaichao committed
Jan 14, 2025 at 19:24 UTC
ff39141a49928540c3975cc8e2a6e9bfa51f0bef
1 file changed
+5
.buildkite/run-hpu-test.sh
+5
@@ -8,6 +8,11 @@ set -ex
8
docker build -t hpu-test-env -f Dockerfile.hpu .
9
10
# Setup cleanup
11
+# certain versions of HPU software stack have a bug that can
12
+# override the exit code of the script, so we need to use
13
+# separate remove_docker_container and remove_docker_container_and_exit
14
+# functions, while other platforms only need one remove_docker_container
15
+# function.
16
EXITCODE=1
17
remove_docker_container() { docker rm -f hpu-test || true; }
18
remove_docker_container_and_exit() { remove_docker_container; exit $EXITCODE; }