Use .par files again

Author: OliverRietmann

.github/workflows/standalone-benchmark.yml

@@ -19,19 +19,19 @@ jobs:
           - name: nvidia-h100
             runner: cern-nextgen-h100
             cmake_args: -DENABLE_CUDA=1 -DENABLE_HIP=0 -DCUDA_COMPUTETARGET=90
-            ca_args: --gpuType CUDA #--RTCTECHloadLaunchBoundsFromFile genGPUArch/nvidia-h100.par
+            ca_args: --gpuType CUDA --RTCTECHloadLaunchBoundsFromFile genGPUArch/nvidia-h100.par
           - name: nvidia-l40s
             runner: cern-nextgen-l40s
             cmake_args: -DENABLE_CUDA=1 -DENABLE_HIP=0 -DCUDA_COMPUTETARGET=89
-            ca_args: --gpuType CUDA #--RTCTECHloadLaunchBoundsFromFile genGPUArch/nvidia-l40s.par
+            ca_args: --gpuType CUDA --RTCTECHloadLaunchBoundsFromFile genGPUArch/nvidia-l40s.par
           - name: amd-mi300x
             runner: cern-nextgen-mi300x
             cmake_args: -DENABLE_CUDA=0 -DENABLE_HIP=1 -DHIP_AMDGPUTARGET=gfx942
-            ca_args: --gpuType HIP #--RTCTECHloadLaunchBoundsFromFile genGPUArch/amd-mi300x.par
+            ca_args: --gpuType HIP --RTCTECHloadLaunchBoundsFromFile genGPUArch/amd-mi300x.par
           - name: amd-w7900
             runner: cern-nextgen-w7900
             cmake_args: -DENABLE_CUDA=0 -DENABLE_HIP=1 -DHIP_AMDGPUTARGET=gfx1100
-            ca_args: --gpuType HIP #--RTCTECHloadLaunchBoundsFromFile genGPUArch/amd-w7900.par
+            ca_args: --gpuType HIP --RTCTECHloadLaunchBoundsFromFile genGPUArch/amd-w7900.par
 
     name: ${{ matrix.name }}
     steps: