[ROCM][CI] Introduce tests-to-include as rocm-test workflow input (#110511)

Fixes https://github.com/pytorch/pytorch/issues/110181

Pull Request resolved: https://github.com/pytorch/pytorch/pull/110511
Approved by: https://github.com/huydhn

.ci/pytorch/test.sh

@@ -80,6 +80,11 @@ if [[ "$BUILD_ENVIRONMENT" != *bazel* ]]; then
   CUSTOM_TEST_ARTIFACT_BUILD_DIR=$(realpath "${CUSTOM_TEST_ARTIFACT_BUILD_DIR:-"build/custom_test_artifacts"}")
 fi
 
+# Reduce set of tests to include when running run_test.py
+if [[ -n $TESTS_TO_INCLUDE ]]; then
+  echo "Setting INCLUDE_CLAUSE"
+  INCLUDE_CLAUSE="--include $TESTS_TO_INCLUDE"
+fi
 
 # shellcheck source=./common.sh
 source "$(dirname "${BASH_SOURCE[0]}")/common.sh"
@@ -228,13 +233,16 @@ test_python_shard() {
     exit 1
   fi
 
-  time python test/run_test.py --exclude-jit-executor --exclude-distributed-tests --shard "$1" "$NUM_TEST_SHARDS" --verbose
+  # Bare --include flag is not supported and quoting for lint ends up with flag not being interpreted correctly
+  # shellcheck disable=SC2086
+  time python test/run_test.py --exclude-jit-executor --exclude-distributed-tests $INCLUDE_CLAUSE --shard "$1" "$NUM_TEST_SHARDS" --verbose
 
   assert_git_not_dirty
 }
 
 test_python() {
-  time python test/run_test.py --exclude-jit-executor --exclude-distributed-tests --verbose
+  # shellcheck disable=SC2086
+  time python test/run_test.py --exclude-jit-executor --exclude-distributed-tests $INCLUDE_CLAUSE --verbose
   assert_git_not_dirty
 }
 
@@ -681,7 +689,8 @@ test_vulkan() {
 
 test_distributed() {
   echo "Testing distributed python tests"
-  time python test/run_test.py --distributed-tests --shard "$SHARD_NUMBER" "$NUM_TEST_SHARDS" --verbose
+  # shellcheck disable=SC2086
+  time python test/run_test.py --distributed-tests --shard "$SHARD_NUMBER" "$NUM_TEST_SHARDS" $INCLUDE_CLAUSE --verbose
   assert_git_not_dirty
 
   if [[ ("$BUILD_ENVIRONMENT" == *cuda* || "$BUILD_ENVIRONMENT" == *rocm*) && "$SHARD_NUMBER" == 1 ]]; then
@@ -1092,6 +1101,10 @@ elif [[ "${BUILD_ENVIRONMENT}" == *-mobile-lightweight-dispatch* ]]; then
   test_libtorch
 elif [[ "${TEST_CONFIG}" = docs_test ]]; then
   test_docs_test
+elif [[ "${BUILD_ENVIRONMENT}" == *rocm* && -n "$TESTS_TO_INCLUDE" ]]; then
+  install_torchvision
+  test_python
+  test_aten
 else
   install_torchvision
   install_monkeytype

.github/workflows/_rocm-test.yml

@@ -32,6 +32,12 @@ on:
         default: 300
         description: |
           Set the maximum (in minutes) how long the workflow should take to finish
+      tests-to-include:
+        required: false
+        type: string
+        default: ""
+        description: |
+          List of tests to include (empty string implies default list)
 
 env:
   GIT_DEFAULT_BRANCH: ${{ github.event.repository.default_branch }}
@@ -136,6 +142,7 @@ jobs:
           XLA_CLANG_CACHE_S3_BUCKET_NAME: ossci-compiler-clang-cache-circleci-xla
           PYTORCH_TEST_CUDA_MEM_LEAK_CHECK: ${{ matrix.mem_leak_check && '1' || '0' }}
           PYTORCH_TEST_RERUN_DISABLED_TESTS: ${{ matrix.rerun_disabled_tests && '1' || '0' }}
+          TESTS_TO_INCLUDE: ${{ inputs.tests-to-include }}
         timeout-minutes: ${{ fromJson(steps.test-timeout.outputs.timeout) }}
         run: |
           set -x
@@ -180,6 +187,7 @@ jobs:
             -e XLA_CLANG_CACHE_S3_BUCKET_NAME \
             -e PYTORCH_TEST_CUDA_MEM_LEAK_CHECK \
             -e PYTORCH_TEST_RERUN_DISABLED_TESTS \
+            -e TESTS_TO_INCLUDE \
             --env-file="/tmp/github_env_${GITHUB_RUN_ID}" \
             --ulimit stack=10485760:83886080 \
             --ulimit core=0 \

.github/workflows/trunk.yml

@@ -174,3 +174,25 @@ jobs:
           { config: "default", shard: 6, num_shards: 6, runner: "windows.g5.4xlarge.nvidia.gpu" },
           { config: "force_on_cpu", shard: 1, num_shards: 1, runner: "windows.4xlarge.nonephemeral" },
         ]}
+
+  linux-focal-rocm5_7-py3_8-build:
+    name: linux-focal-rocm5.7-py3.8
+    uses: ./.github/workflows/_linux-build.yml
+    with:
+      build-environment: linux-focal-rocm5.7-py3.8
+      docker-image-name: pytorch-linux-focal-rocm-n-py3
+      sync-tag: rocm-build
+      test-matrix: |
+        { include: [
+          { config: "default", shard: 1, num_shards: 1, runner: "linux.rocm.gpu" },
+        ]}
+
+  linux-focal-rocm5_7-py3_8-test:
+    name: linux-focal-rocm5.7-py3.8
+    uses: ./.github/workflows/_rocm-test.yml
+    needs: linux-focal-rocm5_7-py3_8-build
+    with:
+      build-environment: linux-focal-rocm5.7-py3.8
+      docker-image: ${{ needs.linux-focal-rocm5_7-py3_8-build.outputs.docker-image }}
+      test-matrix: ${{ needs.linux-focal-rocm5_7-py3_8-build.outputs.test-matrix }}
+      tests-to-include: "test_nn test_torch test_cuda test_ops test_unary_ufuncs test_binary_ufuncs test_autograd"