ci: Move rocm distributed tests to periodic
These jobs have a high TTS and a capacity crunch so moving these to
periodic should resolve both of these issues.
Signed-off-by: Eli Uriegas <eliuriegasfb.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/77205
Approved by: https://github.com/malfet
diff --git a/.github/workflows/periodic.yml b/.github/workflows/periodic.yml
index d52c791..ad3908e 100644
--- a/.github/workflows/periodic.yml
+++ b/.github/workflows/periodic.yml
@@ -84,6 +84,29 @@
AWS_OSSCI_METRICS_V2_ACCESS_KEY_ID: ${{ secrets.AWS_OSSCI_METRICS_V2_ACCESS_KEY_ID }}
AWS_OSSCI_METRICS_V2_SECRET_ACCESS_KEY: ${{ secrets.AWS_OSSCI_METRICS_V2_SECRET_ACCESS_KEY }}
+ linux-bionic-rocm5_1-py3_7-distributed-build:
+ name: linux-bionic-rocm5.1-py3.7-distributed
+ uses: ./.github/workflows/_linux-build.yml
+ with:
+ build-environment: linux-bionic-rocm5.1-py3.7
+ docker-image-name: pytorch-linux-bionic-rocm5.1-py3.7
+
+ linux-bionic-rocm5_1-py3_7-distributed-test:
+ name: linux-bionic-rocm5.1-py3.7-distributed
+ uses: ./.github/workflows/_rocm-test.yml
+ needs: linux-bionic-rocm5_1-py3_7-distributed-build
+ with:
+ build-environment: linux-bionic-rocm5.1-py3.7
+ docker-image: ${{ needs.linux-bionic-rocm5_1-py3_7-distributed-build.outputs.docker-image }}
+ test-matrix: |
+ { include: [
+ { config: "distributed", shard: 1, num_shards: 2, runner: "linux.rocm.gpu" },
+ { config: "distributed", shard: 2, num_shards: 2, runner: "linux.rocm.gpu" },
+ ]}
+ secrets:
+ AWS_OSSCI_METRICS_V2_ACCESS_KEY_ID: ${{ secrets.AWS_OSSCI_METRICS_V2_ACCESS_KEY_ID }}
+ AWS_OSSCI_METRICS_V2_SECRET_ACCESS_KEY: ${{ secrets.AWS_OSSCI_METRICS_V2_SECRET_ACCESS_KEY }}
+
linux-xenial-cuda11_3-py3_7-gcc7-debug-build:
name: linux-xenial-cuda11.3-py3.7-gcc7-debug
uses: ./.github/workflows/_linux-build.yml
diff --git a/.github/workflows/trunk.yml b/.github/workflows/trunk.yml
index ab1c167..935bc33 100644
--- a/.github/workflows/trunk.yml
+++ b/.github/workflows/trunk.yml
@@ -94,29 +94,6 @@
build-environment: linux-xenial-cuda11.3-py3.7-gcc7-no-ops
docker-image-name: pytorch-linux-xenial-cuda11.3-cudnn8-py3-gcc7
- linux-bionic-rocm5_1-py3_7-distributed-build:
- name: linux-bionic-rocm5.1-py3.7-distributed
- uses: ./.github/workflows/_linux-build.yml
- with:
- build-environment: linux-bionic-rocm5.1-py3.7
- docker-image-name: pytorch-linux-bionic-rocm5.1-py3.7
-
- linux-bionic-rocm5_1-py3_7-distributed-test:
- name: linux-bionic-rocm5.1-py3.7-distributed
- uses: ./.github/workflows/_rocm-test.yml
- needs: linux-bionic-rocm5_1-py3_7-distributed-build
- with:
- build-environment: linux-bionic-rocm5.1-py3.7
- docker-image: ${{ needs.linux-bionic-rocm5_1-py3_7-distributed-build.outputs.docker-image }}
- test-matrix: |
- { include: [
- { config: "distributed", shard: 1, num_shards: 2, runner: "linux.rocm.gpu" },
- { config: "distributed", shard: 2, num_shards: 2, runner: "linux.rocm.gpu" },
- ]}
- secrets:
- AWS_OSSCI_METRICS_V2_ACCESS_KEY_ID: ${{ secrets.AWS_OSSCI_METRICS_V2_ACCESS_KEY_ID }}
- AWS_OSSCI_METRICS_V2_SECRET_ACCESS_KEY: ${{ secrets.AWS_OSSCI_METRICS_V2_SECRET_ACCESS_KEY }}
-
pytorch-linux-xenial-py3-clang5-android-ndk-r19c-build:
name: pytorch-linux-xenial-py3-clang5-android-ndk-r19c-build
uses: ./.github/workflows/_android-full-build-test.yml