[ matrixMulDrv (Driver API) ]
> Using CUDA Device [0]: NVIDIA H100 PCIe
> GPU Device has SM 9.0 compute capability
  Total amount of global memory:     85021163520 bytes
> findModulePath found file at <./matrixMul_kernel64.fatbin>
> initCUDA loading module: <./matrixMul_kernel64.fatbin>
> 32 block size selected
Processing time: 0.058000 (ms)
Checking computed result for correctness: Result = PASS

NOTE: The CUDA Samples are not meant for performance measurements. Results may vary when GPU Boost is enabled.