move launch benchmark to third party and add csv parsing

Author: HBN-MichalSzy

.github/workflows/third-party-benchmarks.yml

@@ -223,6 +223,18 @@ jobs:
             --max-new-tokens $MAX_NEW_TOKENS \
             --batch-size $BATCH_SIZE
 
+      - name: Run launch microbenchmark tests
+        if: ${{ steps.install.outcome == 'success' && !cancelled() && (inputs.benchmarks == '' || contains(fromJson(inputs.benchmarks || '[]'), 'launch_micro_benchmarks')) && !contains(fromJson(inputs.skip_benchmarks || '[]'), 'launch_micro_benchmarks') }}
+        run: |
+          source scripts/capture-hw-details.sh
+          python python/test/microbenchmark/launch_overhead.py --reports $REPORTS
+
+          python benchmarks/third_party/vllm/transform_results.py $REPORTS/launch_overhead_results.csv $REPORTS/launch_overhead-report.csv \
+            --tag $TAG \
+            --bgroup overhead \
+            --benchmark launch-overhead \
+            --param_cols="input_type"
+
       - name: Upload benchmark reports
         if: ${{ steps.install.outcome == 'success' && !cancelled() }}
         uses: actions/upload-artifact@v5

.github/workflows/triton-benchmarks.yml

@@ -376,12 +376,6 @@ jobs:
           cd benchmarks/micro_benchmarks
           python run_benchmarks.py --reports $REPORTS
 
-      - name: Run launch microbenchmark tests
-        if: ${{ steps.install.outcome == 'success' && !cancelled() && (inputs.benchmarks == '' || contains(fromJson(inputs.benchmarks || '[]'), 'launch_micro_benchmarks')) && !contains(fromJson(inputs.skip_benchmarks || '[]'), 'launch_micro_benchmarks') }}
-        run: |
-          source scripts/capture-hw-details.sh
-          python python/test/microbenchmark/launch_overhead.py
-
       - name: Upload benchmark reports
         if: ${{ steps.install.outcome == 'success' && !cancelled() }}
         uses: actions/upload-artifact@v5

benchmarks/third_party/vllm/transform_results.py

@@ -46,7 +46,7 @@ def serialize_params(row):
 
     dfs = []
     for compiler_name in compilers:
-        for value_name in ['TFlops', 'GB/s']:
+        for value_name in ['TFlops', 'GB/s', 'time_us']:
             col = f'{compiler_name}-{value_name}'
             if col not in df.columns:
                 continue

python/test/microbenchmark/launch_overhead.py

@@ -99,13 +99,13 @@ def main(use_tensor_desc: bool, reports_dir: str = None):
 
     if reports_dir:
         os.makedirs(reports_dir, exist_ok=True)
-        csv_path = os.path.join(reports_dir, "launch_overhead_report.csv")
+        csv_path = os.path.join(reports_dir, "launch_overhead_results.csv")
         file_exists = os.path.exists(csv_path)
 
         with open(csv_path, "a", newline="") as csvfile:
             writer = csv.writer(csvfile)
             if not file_exists:
-                writer.writerow(["input_type", "median_usecs"])
+                writer.writerow(["input_type", "triton-time_us"])
 
             input_type = "TensorDescriptor" if use_tensor_desc else "Tensor"
             writer.writerow([input_type, round(sorted(usecs)[len(usecs) >> 1], 2)])