Run slow jobs in trunk commits (pytorch#139842)

Per our discussion in https://fburl.com/gdoc/voce5o06, we will run slow jobs more frequently on all trunk commits.  Note that slowgradcheck jobs are moved to periodic as they are not about running slow tests.

There are currently 3 GPU + 2 ROCm + some CPU `linux.4xlarge` runners running slow jobs.  So, I don't expect to see a big increase in CI cost after this.

Also, these slow jobs will only run in trunk commits, not in PRs, so their duration won't affect PR TTS.
Pull Request resolved: pytorch#139842
Approved by: https://github.com/clee2000

Author: huydhn

.github/workflows/periodic.yml

@@ -290,3 +290,36 @@ jobs:
       build-environment: linux-focal-cuda11.8-py3.10-gcc9-experimental-split-build
       docker-image: ${{ needs.linux-focal-cuda11_8-py3_10-gcc9-experimental-split-build.outputs.docker-image }}
       test-matrix: ${{ needs.linux-focal-cuda11_8-py3_10-gcc9-experimental-split-build.outputs.test-matrix }}
+
+  linux-focal-cuda12_1-py3-gcc9-slow-gradcheck-build:
+    name: linux-focal-cuda12.1-py3-gcc9-slow-gradcheck
+    uses: ./.github/workflows/_linux-build.yml
+    needs: get-label-type
+    with:
+      runner_prefix: "${{ needs.get-label-type.outputs.label-type }}"
+      build-environment: linux-focal-cuda12.1-py3-gcc9-slow-gradcheck
+      docker-image-name: pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9
+      cuda-arch-list: 8.6
+      test-matrix: |
+        { include: [
+          { config: "default", shard: 1, num_shards: 8, runner: "linux.g5.4xlarge.nvidia.gpu", owners: ["module:slowgradcheck"] },
+          { config: "default", shard: 2, num_shards: 8, runner: "linux.g5.4xlarge.nvidia.gpu", owners: ["module:slowgradcheck"] },
+          { config: "default", shard: 3, num_shards: 8, runner: "linux.g5.4xlarge.nvidia.gpu", owners: ["module:slowgradcheck"] },
+          { config: "default", shard: 4, num_shards: 8, runner: "linux.g5.4xlarge.nvidia.gpu", owners: ["module:slowgradcheck"] },
+          { config: "default", shard: 5, num_shards: 8, runner: "linux.g5.4xlarge.nvidia.gpu", owners: ["module:slowgradcheck"] },
+          { config: "default", shard: 6, num_shards: 8, runner: "linux.g5.4xlarge.nvidia.gpu", owners: ["module:slowgradcheck"] },
+          { config: "default", shard: 7, num_shards: 8, runner: "linux.g5.4xlarge.nvidia.gpu", owners: ["module:slowgradcheck"] },
+          { config: "default", shard: 8, num_shards: 8, runner: "linux.g5.4xlarge.nvidia.gpu", owners: ["module:slowgradcheck"] },
+        ]}
+
+  linux-focal-cuda12_1-py3-gcc9-slow-gradcheck-test:
+    name: linux-focal-cuda12.1-py3-gcc9-slow-gradcheck
+    uses: ./.github/workflows/_linux-test.yml
+    needs:
+      - linux-focal-cuda12_1-py3-gcc9-slow-gradcheck-build
+      - target-determination
+    with:
+      build-environment: linux-focal-cuda12.1-py3-gcc9-slow-gradcheck
+      docker-image: ${{ needs.linux-focal-cuda12_1-py3-gcc9-slow-gradcheck-build.outputs.docker-image }}
+      test-matrix: ${{ needs.linux-focal-cuda12_1-py3-gcc9-slow-gradcheck-build.outputs.test-matrix }}
+      timeout-minutes: 300

.github/workflows/slow.yml

@@ -4,14 +4,14 @@
 name: slow
 
 on:
-  schedule:
-    - cron: 45 0,4,8,12,16,20 * * *
-    - cron: 29 8 * * *  # about 1:29am PDT, for mem leak check and rerun disabled tests
   push:
-    tags:
-      - ciflow/slow/*
     branches:
+      - main
       - release/*
+    tags:
+      - ciflow/slow/*
+  schedule:
+    - cron: 29 8 * * *  # about 1:29am PDT, for mem leak check and rerun disabled tests
   workflow_dispatch:
 
 concurrency:
@@ -47,39 +47,6 @@ jobs:
       curr_branch: ${{ github.head_ref || github.ref_name }}
       curr_ref_type: ${{ github.ref_type }}
 
-  linux-focal-cuda12_1-py3-gcc9-slow-gradcheck-build:
-    name: linux-focal-cuda12.1-py3-gcc9-slow-gradcheck
-    uses: ./.github/workflows/_linux-build.yml
-    needs: get-label-type
-    with:
-      runner_prefix: "${{ needs.get-label-type.outputs.label-type }}"
-      build-environment: linux-focal-cuda12.1-py3-gcc9-slow-gradcheck
-      docker-image-name: pytorch-linux-focal-cuda12.1-cudnn9-py3-gcc9
-      cuda-arch-list: 8.6
-      test-matrix: |
-        { include: [
-          { config: "default", shard: 1, num_shards: 8, runner: "linux.g5.4xlarge.nvidia.gpu", owners: ["module:slowgradcheck"] },
-          { config: "default", shard: 2, num_shards: 8, runner: "linux.g5.4xlarge.nvidia.gpu", owners: ["module:slowgradcheck"] },
-          { config: "default", shard: 3, num_shards: 8, runner: "linux.g5.4xlarge.nvidia.gpu", owners: ["module:slowgradcheck"] },
-          { config: "default", shard: 4, num_shards: 8, runner: "linux.g5.4xlarge.nvidia.gpu", owners: ["module:slowgradcheck"] },
-          { config: "default", shard: 5, num_shards: 8, runner: "linux.g5.4xlarge.nvidia.gpu", owners: ["module:slowgradcheck"] },
-          { config: "default", shard: 6, num_shards: 8, runner: "linux.g5.4xlarge.nvidia.gpu", owners: ["module:slowgradcheck"] },
-          { config: "default", shard: 7, num_shards: 8, runner: "linux.g5.4xlarge.nvidia.gpu", owners: ["module:slowgradcheck"] },
-          { config: "default", shard: 8, num_shards: 8, runner: "linux.g5.4xlarge.nvidia.gpu", owners: ["module:slowgradcheck"] },
-        ]}
-
-  linux-focal-cuda12_1-py3-gcc9-slow-gradcheck-test:
-    name: linux-focal-cuda12.1-py3-gcc9-slow-gradcheck
-    uses: ./.github/workflows/_linux-test.yml
-    needs:
-      - linux-focal-cuda12_1-py3-gcc9-slow-gradcheck-build
-      - target-determination
-    with:
-      build-environment: linux-focal-cuda12.1-py3-gcc9-slow-gradcheck
-      docker-image: ${{ needs.linux-focal-cuda12_1-py3-gcc9-slow-gradcheck-build.outputs.docker-image }}
-      test-matrix: ${{ needs.linux-focal-cuda12_1-py3-gcc9-slow-gradcheck-build.outputs.test-matrix }}
-      timeout-minutes: 300
-
   linux-focal-cuda12_1-py3_10-gcc9-sm86-build:
     name: linux-focal-cuda12.1-py3.10-gcc9-sm86
     uses: ./.github/workflows/_linux-build.yml