various updates to "build whl" workflow (#59)

SUMMARY:
* updates whl generation workflow to add testing and `testmo`
integration
* add top-level generate whls workflow

TEST PLAN:
ran manually ...

---------

Co-authored-by: andy-neuma <[email protected]>
Co-authored-by: Varun Sundar Rabindranath <[email protected]>
Co-authored-by: Michael Goin <[email protected]>

Author: 4 people

.github/actions/nm-build-vllm-whl/action.yml

@@ -13,7 +13,7 @@ outputs:
     value: ${{ steps.whl.outputs.status }}
   whl:
     description: 'basename for generated whl'
-    value: ${{ steps.build_whl.outputs.whl }}
+    value: ${{ steps.whl.outputs.whl }}
 runs:
   using: composite
   steps:
@@ -25,6 +25,7 @@ runs:
       SUCCESS=0
       pip3 wheel --no-deps -w dist . || SUCCESS=$?
       echo "status=${SUCCESS}" >> "$GITHUB_OUTPUT"
+      ls -alh dist/
       BASE=$(./.github/scripts/convert-version ${{ inputs.python }})
       WHL_FILEPATH=$(find dist -iname "*nm_vllm*${BASE}*.whl")
       WHL=$(basename ${WHL_FILEPATH})

.github/actions/nm-install-test-whl/action.yml

@@ -0,0 +1,40 @@
+name: install and test whl
+description: 'install whl in venv and run tests'
+inputs:
+  python:
+    description: 'python version, e.g. 3.10.12'
+    required: true
+  venv:
+    description: 'name for python virtual environment'
+    required: true
+  whl:
+    description: 'filepath for magic_wand whl file'
+    required: true
+  test_directory:
+    description: 'directory to test files'
+    required: true
+  test_results:
+    description: 'desired directory for test results'
+    required: true
+outputs:
+  status:
+    description: 'return test result'
+    value: ${{ steps.whl_install_test.outputs.status }}
+runs:
+  using: composite
+  steps:
+    - id: whl_install_test
+      run: |
+        COMMIT=${{ github.sha }}
+        VENV="${{ inputs.venv }}-${COMMIT:0:7}"
+        pyenv virtualenv ${VENV}
+        source $(pyenv root)/versions/${{ inputs.python }}/envs/${VENV}/bin/activate
+        pip3 list
+        ls -alh dist
+        pip3 install ${{ inputs.whl }}
+        pip3 install -r requirements-dev.txt
+        SUCCESS=0
+        ./.github/scripts/run-tests -t ${{ inputs.test_directory }} -r ${{ inputs.test_results }} || SUCCESS=$?
+        echo "status=${SUCCESS}" >> "$GITHUB_OUTPUT"
+        exit ${SUCCESS}
+      shell: bash

.github/actions/nm-set-env/action.yml

@@ -11,6 +11,7 @@ runs:
   using: composite
   steps:
   - run: |
+      echo "TORCH_CUDA_ARCH_LIST=7.0 7.5 8.0 8.6 8.9 9.0+PTX" >> $GITHUB_ENV
       echo "HF_TOKEN=${HF_TOKEN_SECRET}" >> $GITHUB_ENV
       echo "HF_HOME=/EFS/hf_home" >> $GITHUB_ENV
       NUM_THREADS=$(./.github/scripts/determine-threading -G ${{ inputs.Gi_per_thread }})

.github/actions/nm-test-vllm/action.yml

@@ -30,7 +30,7 @@ runs:
       source $(pyenv root)/versions/${{ inputs.python }}/envs/${VENV}/bin/activate
       pip3 install --index-url http://${{ inputs.pypi }}:8080/ --trusted-host ${{ inputs.pypi }} magic-wand
       pip3 install -r requirements-dev.txt
-      # run tests via runner script (serially)
+      # run tests via runner script
       SUCCESS=0
       ./.github/scripts/run-tests -t ${{ inputs.test_directory }} -r ${{ inputs.test_results }} || SUCCESS=$?
       echo "was this a SUCCESS? ${SUCCESS}"

.github/actions/nm-whl-summary/action.yml

@@ -0,0 +1,37 @@
+name: summary whl
+description: 'creates a summary for whl workflow'
+inputs:
+  label:
+    description: 'GHA runner label'
+    required: true
+  gitref:
+    description: 'git commit hash or branch name'
+    required: true
+  testmo_run_url:
+    description: 'testmo URL for this particular run'
+    required: true
+  python:
+    description: 'python version info'
+    required: true
+  whl:
+    description: 'whl file tested'
+    required: true
+runs:
+  using: composite
+  steps:
+  - run: |
+      TESTMO_URL=${{ inputs.testmo_run_url }}
+      TEST_STATUS=${{ inputs.test_status }}
+      TEST_EMOJI=$(./.github/scripts/step-status ${TEST_STATUS})
+      echo "testmo URL: ${TESTMO_URL}" >> $GITHUB_STEP_SUMMARY
+      echo ""
+      echo "| Parameter | |" >> $GITHUB_STEP_SUMMARY
+      echo "|---|---|" >> $GITHUB_STEP_SUMMARY
+      echo "| label: | \`${{ inputs.label }}\` |" >> $GITHUB_STEP_SUMMARY
+      echo "| git sha: | \`${{ github.sha }}\` |" >> $GITHUB_STEP_SUMMARY
+      echo "| github actor: | '${{ github.actor }}' |" >> $GITHUB_STEP_SUMMARY
+      echo "| gitref: | '${{ inputs.gitref }}' |" >> $GITHUB_STEP_SUMMARY
+      echo "| branch name: | '${{ github.ref_name }}' |" >> $GITHUB_STEP_SUMMARY
+      echo "| python: | ${{ inputs.python }} |" >> $GITHUB_STEP_SUMMARY
+      echo "| whl: | ${{ inputs.whl }} |" >> $GITHUB_STEP_SUMMARY
+    shell: bash

.github/workflows/build-whl.yml

@@ -49,7 +49,8 @@ on:
         required: true
 
 env:
-    VENV_BASE: "WHL-TEST"
+    VENV_BUILD_BASE: "BUILD"
+    VENV_WHL_BASE: "WHL-TEST"
     WHL_TEST_RESULTS: "whl-test-results"
 
 jobs:
@@ -81,7 +82,7 @@ jobs:
               uses: ./.github/actions/nm-set-python/
               with:
                 python: ${{ inputs.python }}
-                venv: ${{ env.VENV_BASE }}
+                venv: ${{ env.VENV_BUILD_BASE }}
 
             - name: create testmo run
               id: create_testmo_run
@@ -102,14 +103,15 @@ jobs:
               with:
                 Gi_per_thread: ${{ inputs.Gi_per_thread }}
                 python: ${{ inputs.python }}
-                venv: ${{ env.VENV_BASE }}
+                venv: ${{ env.VENV_BUILD_BASE }}
                 pypi: ${{ secrets.NM_PRIVATE_PYPI_LOCATION }}
 
             - name: build whl
               id: build_whl
               uses: ./.github/actions/nm-build-vllm-whl/
               with:
                 python: ${{ inputs.python }}
+                venv: ${{ env.VENV_BUILD_BASE }}
 
             - name: upload whl
               uses: actions/upload-artifact@v4
@@ -119,8 +121,17 @@ jobs:
                 path: dist/${{ steps.build_whl.outputs.whl }}
                 retention-days: 15
 
+            - name: summary
+              uses: ./.github/actions/nm-whl-summary/
+              with:
+                label: ${{ inputs.build_label }}
+                gitref: ${{ inputs.gitref }}
+                testmo_run_url: https://neuralmagic.testmo.net/automation/runs/view/${{ steps.create_testmo_run.outputs.id }}
+                python: ${{ inputs.python }}
+                whl: ${{ steps.build_whl.outputs.whl }}
+
             - name: complete testmo run
-              uses: ./.github/actions/testmo-run-complete/
+              uses: ./.github/actions/nm-testmo-run-complete/
               if: success() || failure()
               with:
                 testmo_url: https://neuralmagic.testmo.net

.github/workflows/gen-whl.yml

@@ -0,0 +1,24 @@
+name: generate whl
+run-name: ${{ github.actor }} generating whl on branch '${{ github.ref }}'
+on:
+  workflow_dispatch:
+    inputs:
+        gitref:
+            description: 'git commit hash or branch name'
+            type: string
+            required: true
+
+jobs:
+
+  AWS-AVX2-32G-A10G-24G:
+    strategy:
+      matrix:
+            python: [3.8.17, 3.9.17, 3.10.12, 3.11.4]
+    uses: ./.github/workflows/build-whl.yml
+    with:
+        build_label: aws-avx2-192G-4-a10g-96G
+        timeout: 30
+        gitref: ${{ inputs.gitref }}
+        Gi_per_thread: 4
+        python: ${{ matrix.python }}
+    secrets: inherit

.github/workflows/remote-push.yml

@@ -37,6 +37,6 @@ jobs:
             label: aws-avx2-32G-a10g-24G
             timeout: 180
             gitref: '${{ github.ref }}'
-            Gi_per_thread: 4
+            Gi_per_thread: 12
             python: ${{ matrix.python }}
         secrets: inherit