[fbsync] enable Windows GPU CI on GHA (#7498)

Reviewed By: vmoens

Differential Revision: D45522826

fbshipit-source-id: fb832f29c09ade1d7a796ebf31d35d74ed600aff

Author: NicolasHug

.circleci/config.yml

@@ -433,53 +433,6 @@ jobs:
             docker push ${image_name}:${CIRCLE_WORKFLOW_ID}
             docker push ${image_name}:latest
 
-  unittest_windows_gpu:
-    <<: *binary_common
-    executor:
-      name: windows-gpu
-    environment:
-      CUDA_VERSION: "11.7"
-      PYTHON_VERSION: << parameters.python_version >>
-    steps:
-      - checkout
-      - designate_upload_channel
-      - run:
-          name: Generate cache key
-          # This will refresh cache on Sundays, nightly build should generate new cache.
-          command: echo "$(date +"%Y-%U")" > .circleci-weekly
-      - restore_cache:
-          {% raw %}
-          keys:
-            - env-v1-windows-{{ arch }}-py<< parameters.python_version >>-{{ checksum ".circleci/unittest/windows/scripts/environment.yml" }}-{{ checksum ".circleci-weekly" }}
-          {% endraw %}
-      - run:
-          name: Setup
-          command: .circleci/unittest/windows/scripts/setup_env.sh
-      - save_cache:
-          {% raw %}
-          key: env-v1-windows-{{ arch }}-py<< parameters.python_version >>-{{ checksum ".circleci/unittest/windows/scripts/environment.yml" }}-{{ checksum ".circleci-weekly" }}
-          {% endraw %}
-          paths:
-            - conda
-            - env
-      - run:
-          name: Install CUDA
-          command: packaging/windows/internal/cuda_install.bat
-      - run:
-          name: Update CUDA driver
-          command: packaging/windows/internal/driver_update.bat
-      - run:
-          name: Install torchvision
-          command: .circleci/unittest/windows/scripts/install.sh
-      - run:
-          name: Run tests
-          command: .circleci/unittest/windows/scripts/run_test.sh
-      - run:
-          name: Post process
-          command: .circleci/unittest/windows/scripts/post_process.sh
-      - store_test_results:
-          path: test-results
-
   cmake_linux_cpu:
     <<: *binary_common
     docker:
@@ -575,10 +528,6 @@ workflows:
       {{ ios_workflows() }}
       {{ android_workflows() }}
 
-  unittest:
-    jobs:
-      {{ unittest_workflows() }}
-
   cmake:
     jobs:
       {{ cmake_workflows() }}

.circleci/config.yml.in

@@ -217,36 +217,6 @@ def indent(indentation, data_list):
     return ("\n" + " " * indentation).join(yaml.dump(data_list, default_flow_style=False).splitlines())
 
 
-def unittest_workflows(indentation=6):
-    jobs = []
-    for os_type in ["windows"]:
-        for device_type in ["gpu"]:
-            if os_type == "macos" and device_type == "gpu":
-                continue
-
-            for i, python_version in enumerate(PYTHON_VERSIONS):
-
-                # Turn off unit tests for 3.11, unit test are not setup properly
-                if python_version == "3.11":
-                    continue
-
-                job = {
-                    "name": f"unittest_{os_type}_{device_type}_py{python_version}",
-                    "python_version": python_version,
-                }
-
-                if device_type == "gpu":
-                    if python_version != "3.8":
-                        job["filters"] = gen_filter_branch_tree("main", "nightly")
-                    job["cu_version"] = "cu117"
-                else:
-                    job["cu_version"] = "cpu"
-
-                jobs.append({f"unittest_{os_type}_{device_type}": job})
-
-    return indent(indentation, jobs)
-
-
 def cmake_workflows(indentation=6):
     jobs = []
     python_version = "3.8"
@@ -331,7 +301,6 @@ def android_workflows(indentation=6, nightly=False):
         f.write(
             env.get_template("config.yml.in").render(
                 build_workflows=build_workflows,
-                unittest_workflows=unittest_workflows,
                 cmake_workflows=cmake_workflows,
                 ios_workflows=ios_workflows,
                 android_workflows=android_workflows,

.circleci/regenerate.py

@@ -20,12 +20,11 @@ jobs:
           - "3.11"
         runner: ["windows.4xlarge"]
         gpu-arch-type: ["cpu"]
-        # FIXME: enable this as soon as nvjpeg is available on the Windows runner
-#        include:
-#          - python-version: "3.8"
-#            runner: windows.8xlarge.nvidia.gpu
-#            gpu-arch-type: cuda
-#            gpu-arch-version: "11.7"
+        include:
+          - python-version: "3.8"
+            runner: windows.g5.4xlarge.nvidia.gpu
+            gpu-arch-type: cuda
+            gpu-arch-version: "11.7"
       fail-fast: false
     uses: pytorch/test-infra/.github/workflows/windows_job.yml@main
     with:
@@ -46,6 +45,7 @@ jobs:
           # TODO: This should be handled by the generic Windows job the same as its done by the generic Linux job
           export CUDA_HOME="/c/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v${{ matrix.gpu-arch-version }}"
           export CUDA_PATH="${CUDA_HOME}"
+          export PATH="${CUDA_PATH}/bin:${PATH}"
         fi
         
         ./.github/scripts/unittest.sh