diff --git a/TPs/TP1/CODE/Partie2/dgemm.cu b/TPs/TP1/CODE/Partie2/dgemm.cu index d31c37d6e137765420577c3c4e7ca991696eef9e..9bc64e9eef89414afa88e83a593fa25a2cbd7845 100755 --- a/TPs/TP1/CODE/Partie2/dgemm.cu +++ b/TPs/TP1/CODE/Partie2/dgemm.cu @@ -282,19 +282,19 @@ int main(int argc, char** argv){ - cudaEventRecord(ct0); - d_shared_mult<<<dimGrid, dimBlock>>>(A_d, B_d, C_d, N); - cudaEventRecord(ct1); - cudaDeviceSynchronize(); - getLastCudaError("Mult kernel Anzo failed"); - cudaEventElapsedTime(&cudaDuration, ct0, ct1); - checkCudaErrors(cudaMemcpy(C_h, C_d, - sizeof(double) * N * N, cudaMemcpyDeviceToHost)); - - fprintf(stdout, "GPU Performance results Anzo: \n"); - fprintf(stdout, " Time: %lf s\n", cudaDuration / 1000); - fprintf(stdout, " MFlops: %.2f\n", (nb_op / cudaDuration) * 1E-3); - verify_matrix(C_cpu, C_h, N); +// cudaEventRecord(ct0); +// d_shared_mult<<<dimGrid, dimBlock>>>(A_d, B_d, C_d, N); +// cudaEventRecord(ct1); +// cudaDeviceSynchronize(); +// getLastCudaError("Mult kernel Anzo failed"); +// cudaEventElapsedTime(&cudaDuration, ct0, ct1); +// checkCudaErrors(cudaMemcpy(C_h, C_d, +// sizeof(double) * N * N, cudaMemcpyDeviceToHost)); +// +// fprintf(stdout, "GPU Performance results Anzo: \n"); +// fprintf(stdout, " Time: %lf s\n", cudaDuration / 1000); +// fprintf(stdout, " MFlops: %.2f\n", (nb_op / cudaDuration) * 1E-3); +// verify_matrix(C_cpu, C_h, N); return 0; }