[rocprofiler-systems] Add different test coverage for CI/Nightly, add better logging for failures

.github/workflows/rocprofiler-systems-continuous-integration.yml

@@ -36,18 +36,34 @@ env:
   ROCM_VERSION: "7.0.0"
 
 jobs:
+  prepare_matrix:
+    runs-on: ubuntu-latest
+    outputs:
+      matrix: ${{ steps.generate_matrix.outputs.matrix }}
+
+    steps:
+      - uses: actions/checkout@v5
+        with:
+          sparse-checkout: projects/rocprofiler-systems/.github
+
+      - name: Generate and output matrix
+        id: generate_matrix
+        working-directory: projects/rocprofiler-systems/.github
+        run: |
+          if [ '${{ github.event_name }}' = 'schedule' ]; then
+            MATRIX_CONTENT=$(cat ci-matrix.yml | yq '.matrix-nightly' -I=0 -o=json)
+          else
+            MATRIX_CONTENT=$(cat ci-matrix.yml | yq '.matrix-ci' -I=0 -o=json)
+          fi
+          echo "matrix=${MATRIX_CONTENT}" >> $GITHUB_OUTPUT
+
   ubuntu:
     name: Ubuntu ${{ matrix.system.os-release }} • ${{ matrix.system.arch }}
+    needs: prepare_matrix
     strategy:
       fail-fast: false
       matrix:
-        system:
-          # MI355
-          - { os-release: '22.04', gpu: 'mi355', arch: 'gfx950', runner: 'linux-mi355-1gpu-ossci-rocm' }
-          - { os-release: '24.04', gpu: 'mi355', arch: 'gfx950', runner: 'linux-mi355-1gpu-ossci-rocm' }
-          # MI325
-          - { os-release: '22.04', gpu: 'mi325', arch: 'gfx94X', runner: 'linux-mi325-1gpu-ossci-rocm' }
-          - { os-release: '24.04', gpu: 'mi325', arch: 'gfx94X', runner: 'linux-mi325-1gpu-ossci-rocm' }
+        system: ${{ fromJSON(needs.prepare_matrix.outputs.matrix) }}
     runs-on: ${{ matrix.system.runner }}
     env:
       HIP_PLATFORM: "amd"
@@ -65,6 +81,7 @@ jobs:
         --device /dev/kfd
         --device /dev/dri
         --cap-add CAP_SYS_ADMIN
+
     steps:
       - uses: actions/checkout@v5
         with:
@@ -122,6 +139,7 @@ jobs:
           echo "✅ ROCm Installation Complete!"
 
       - name: Configure, Build, and Test
+        id: run_ci
         timeout-minutes: 30
         working-directory: projects/rocprofiler-systems
         run: |
@@ -152,3 +170,10 @@ jobs:
             -- \
             -L "rocm|network" \
             -LE "rccl|runtime|ompvv"
+
+      - name: Output Logs
+        if: failure() && steps.run_ci.outcome == 'failure'
+        working-directory: projects/rocprofiler-systems
+        run: |
+          echo "❌ Run Failed: Now outputting LastTest.log files for detailed logs..."
+          cat build/Testing/Temporary/LastTest*.log

projects/rocprofiler-systems/.github/ci-matrix.yml

@@ -0,0 +1,12 @@
+# Matrix definitions for rocprofiler-systems-continuous-integration
+
+matrix-nightly:
+  # MI355
+  - { os-release: '22.04', gpu: 'mi355', arch: 'gfx950', runner: 'linux-mi355-1gpu-ossci-rocm' }
+  - { os-release: '24.04', gpu: 'mi355', arch: 'gfx950', runner: 'linux-mi355-1gpu-ossci-rocm' }
+  # MI325
+  - { os-release: '22.04', gpu: 'mi325', arch: 'gfx94X', runner: 'linux-mi325-1gpu-ossci-rocm' }
+  - { os-release: '24.04', gpu: 'mi325', arch: 'gfx94X', runner: 'linux-mi325-1gpu-ossci-rocm' }
+matrix-ci:
+  # MI355
+  - { os-release: '24.04', gpu: 'mi355', arch: 'gfx950', runner: 'linux-mi355-1gpu-ossci-rocm' }