Unverified Commit 1bd8a72f authored by Frank Lee's avatar Frank Lee Committed by GitHub
Browse files

[workflow] disable p2p via shared memory on non-nvlink machine (#1086)

parent 65ee6dcc
...@@ -15,7 +15,7 @@ jobs: ...@@ -15,7 +15,7 @@ jobs:
runs-on: [self-hosted, gpu] runs-on: [self-hosted, gpu]
container: container:
image: hpcaitech/pytorch-cuda:1.10.1-11.3.0 image: hpcaitech/pytorch-cuda:1.10.1-11.3.0
options: --shm-size=2gb --gpus all --rm -v /data/scratch/cifar-10:/data/scratch/cifar-10 options: --gpus all --rm -v /data/scratch/cifar-10:/data/scratch/cifar-10
timeout-minutes: 40 timeout-minutes: 40
steps: steps:
- uses: actions/checkout@v2 - uses: actions/checkout@v2
...@@ -34,3 +34,4 @@ jobs: ...@@ -34,3 +34,4 @@ jobs:
PYTHONPATH=$PWD pytest tests PYTHONPATH=$PWD pytest tests
env: env:
DATA: /data/scratch/cifar-10 DATA: /data/scratch/cifar-10
NCCL_SHM_DISABLE: 1
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment