Repository navigation
Expand file tree
/
Copy pathmain.cu
More file actions
108 lines (90 loc) · 3.36 KB
/
Copy pathmain.cu
File metadata and controls
108 lines (90 loc) · 3.36 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
#include <iostream>
#include <chrono>
#include <cuda_runtime.h>
// Matrix dimensions
const int N = 1024; // Matrix size N x N
// CUDA kernel for matrix multiplication
__global__ void matrixMultiplyKernel(float* A, float* B, float* C, int n) {
int row = blockIdx.y * blockDim.y + threadIdx.y;
int col = blockIdx.x * blockDim.x + threadIdx.x;
if (row < n && col < n) {
float sum = 0.0f;
for (int k = 0; k < n; k++) {
sum += A[row * n + k] * B[k * n + col];
}
C[row * n + col] = sum;
}
}
// CPU matrix multiplication
void matrixMultiplyCPU(float* A, float* B, float* C, int n) {
for (int i = 0; i < n; i++) {
for (int j = 0; j < n; j++) {
float sum = 0.0f;
for (int k = 0; k < n; k++) {
sum += A[i * n + k] * B[k * n + j];
}
C[i * n + j] = sum;
}
}
}
// Initialize matrix with random values
void initializeMatrix(float* matrix, int n) {
for (int i = 0; i < n * n; i++) {
matrix[i] = static_cast<float>(rand()) / RAND_MAX;
}
}
int main() {
float *h_A, *h_B, *h_C_cpu, *h_C_gpu; // Host matrices
float *d_A, *d_B, *d_C; // Device matrices
// Allocate host memory
h_A = new float[N * N];
h_B = new float[N * N];
h_C_cpu = new float[N * N];
h_C_gpu = new float[N * N];
// Initialize matrices
initializeMatrix(h_A, N);
initializeMatrix(h_B, N);
// CPU Matrix Multiplication
auto cpu_start = std::chrono::high_resolution_clock::now();
matrixMultiplyCPU(h_A, h_B, h_C_cpu, N);
auto cpu_end = std::chrono::high_resolution_clock::now();
auto cpu_duration = std::chrono::duration_cast<std::chrono::milliseconds>(cpu_end - cpu_start);
// Allocate device memory
cudaMalloc(&d_A, N * N * sizeof(float));
cudaMalloc(&d_B, N * N * sizeof(float));
cudaMalloc(&d_C, N * N * sizeof(float));
// Copy data to device
cudaMemcpy(d_A, h_A, N * N * sizeof(float), cudaMemcpyHostToDevice);
cudaMemcpy(d_B, h_B, N * N * sizeof(float), cudaMemcpyHostToDevice);
// Define grid and block dimensions
dim3 blockDim(16, 16);
dim3 gridDim((N + blockDim.x - 1) / blockDim.x,
(N + blockDim.y - 1) / blockDim.y);
// GPU Matrix Multiplication
auto gpu_start = std::chrono::high_resolution_clock::now();
matrixMultiplyKernel<<<gridDim, blockDim>>>(d_A, d_B, d_C, N);
cudaDeviceSynchronize();
auto gpu_end = std::chrono::high_resolution_clock::now();
auto gpu_duration = std::chrono::duration_cast<std::chrono::milliseconds>(gpu_end - gpu_start);
// Copy result back to host
cudaMemcpy(h_C_gpu, d_C, N * N * sizeof(float), cudaMemcpyDeviceToHost);
// Verify results
float maxError = 0.0f;
for (int i = 0; i < N * N; i++) {
maxError = std::max(maxError, std::abs(h_C_cpu[i] - h_C_gpu[i]));
}
// Print results
std::cout << "Matrix size: " << N << " x " << N << std::endl;
std::cout << "CPU Time: " << cpu_duration.count() << " ms" << std::endl;
std::cout << "GPU Time: " << gpu_duration.count() << " ms" << std::endl;
std::cout << "Maximum Error: " << maxError << std::endl;
// Cleanup
delete[] h_A;
delete[] h_B;
delete[] h_C_cpu;
delete[] h_C_gpu;
cudaFree(d_A);
cudaFree(d_B);
cudaFree(d_C);
return 0;
}