matrix_mul.hh
No OneTemporary
Actions

Subscribers

None

File Metadata

	/*
	* Copyright 1993-2009 NVIDIA Corporation. All rights reserved.
	*
	* NVIDIA Corporation and its licensors retain all intellectual property and
	* proprietary rights in and to this software and related documentation and
	* any modifications thereto. Any use, reproduction, disclosure, or
	* distribution of this software and related documentation without an express
	* license agreement from NVIDIA Corporation is strictly prohibited.
	*
	*/
	#include <cuda_runtime.h>
	#include <iostream>
	#include <random>

	#include "matrix_mul_cpu.hh"
	#include "matrix_mul_gpu.hh"

	#ifndef MATRIX_MUL_HH
	#define MATRIX_MUL_HH

	#define CHECK_RESULT 1

	inline void printDiff(const matrix_cpu &host_A, const matrix_gpu &device_A) {
	int error_count = 0;
	for (int i = 0; i < host_A.rows(); i++) {
	for (int j = 0; j < host_A.cols(); j++) {
	if (std::abs(host_A(i, j) - device_A(i, j)) > 0.1) {
	std::cout << "diff(" << i << ", " << j << ") CPU=" << host_A(i, j)
	<< ", GPU=" << device_A(i, j) << std::endl;
	error_count++;
	}
	}
	}
	std::cout << "Total Errors = " << error_count << std::endl;
	}

	// Thread block size
	#define BLOCK_SIZE 32

	// Matrix dimensions
	// (chosen as multiples of the thread block size for simplicity)
	#define WA (32 * BLOCK_SIZE) // Matrix A width
	#define HA (16 * BLOCK_SIZE) // Matrix A height
	#define WB (24 * BLOCK_SIZE) // Matrix B width
	#define HB WA // Matrix B height
	#define WC WB // Matrix C width
	#define HC HA // Matrix C height

	#endif // MATRIX_MUL_HH