[CI] Add opt-in h100 tests (#153170)

So far only run:
 - inductor/test_fp8.py
 - test_matmul_cuda.py
 - inductor/test_max_autotune.py
Pull Request resolved: https://github.com/pytorch/pytorch/pull/153170
Approved by: https://github.com/drisspg, https://github.com/eellison

.ci/pytorch/test.sh

@@ -314,6 +314,12 @@ test_python() {
   assert_git_not_dirty
 }
 
+test_python_smoke() {
+  # Smoke tests for H100
+  time python test/run_test.py --include test_matmul_cuda inductor/test_fp8 inductor/test_max_autotune $PYTHON_TEST_EXTRA_OPTION --upload-artifacts-while-running
+  assert_git_not_dirty
+}
+
 test_lazy_tensor_meta_reference_disabled() {
   export TORCH_DISABLE_FUNCTIONALIZATION_META_REFERENCE=1
   echo "Testing lazy tensor operations without meta reference"
@@ -1706,6 +1712,8 @@ elif [[ "${BUILD_ENVIRONMENT}" == *xpu* ]]; then
   test_python
   test_aten
   test_xpu_bin
+elif [[ "${TEST_CONFIG}" == smoke ]]; then
+  test_python_smoke
 else
   install_torchvision
   install_monkeytype

.github/pytorch-probot.yml

@@ -28,6 +28,7 @@ ciflow_push_tags:
 - ciflow/autoformat
 - ciflow/op-benchmark
 - ciflow/pull
+- ciflow/h100
 retryable_workflows:
 - pull
 - trunk

.github/workflows/test-h100.yml

@@ -0,0 +1,53 @@
+name: Limited CI on H100
+
+on:
+  pull_request:
+    paths:
+      - .github/workflows/test-h100.yml
+  workflow_dispatch:
+  push:
+    tags:
+      - ciflow/h100/*
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.sha }}-${{ github.event_name == 'workflow_dispatch' }}-${{ github.event_name == 'schedule' }}
+  cancel-in-progress: true
+
+jobs:
+
+  get-label-type:
+    if: github.repository_owner == 'pytorch'
+    name: get-label-type
+    uses: pytorch/pytorch/.github/workflows/_runner-determinator.yml@main
+    with:
+      triggering_actor: ${{ github.triggering_actor }}
+      issue_owner: ${{ github.event.pull_request.user.login || github.event.issue.user.login }}
+      curr_branch: ${{ github.head_ref || github.ref_name }}
+      curr_ref_type: ${{ github.ref_type }}
+
+  linux-focal-cuda12_6-py3_10-gcc11-sm90-build:
+    name: linux-focal-cuda12.6-py3.10-gcc11-sm90
+    uses: ./.github/workflows/_linux-build.yml
+    needs: get-label-type
+    with:
+      runner_prefix: "${{ needs.get-label-type.outputs.label-type }}"
+      runner: "linux.12xlarge"
+      build-environment: linux-focal-cuda12.6-py3.10-gcc11-sm90
+      docker-image-name: ci-image:pytorch-linux-focal-cuda12.6-cudnn9-py3-gcc11
+      cuda-arch-list: '9.0'
+      test-matrix: |
+        { include: [
+          { config: "smoke", shard: 1, num_shards: 1, runner: "linux.aws.h100" },
+        ]}
+    secrets: inherit
+
+  linux-focal-cuda12_6-py3_10-gcc11-sm90-test:
+    name: linux-focal-cuda12.6-py3.10-gcc11-sm90
+    uses: ./.github/workflows/_linux-test.yml
+    needs:
+      - linux-focal-cuda12_6-py3_10-gcc11-sm90-build
+    with:
+      build-environment: linux-focal-cuda12.6-py3.10-gcc11-sm90
+      docker-image: ${{ needs.linux-focal-cuda12_6-py3_10-gcc11-sm90-build.outputs.docker-image }}
+      test-matrix: ${{ needs.linux-focal-cuda12_6-py3_10-gcc11-sm90-build.outputs.test-matrix }}
+    secrets: inherit