array_multiply.cuh
2.97 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
109
110
111
112
113
114
115
116
117
118
119
120
#ifndef STIM_CUDA_ARRAY_MULTIPLY_H
#define STIM_CUDA_ARRAY_MULTIPLY_H
#include <iostream>
#include <cuda.h>
#include <stim/cuda/cudatools.h>
namespace stim{
namespace cuda{
template<typename T>
__global__ void cuda_multiply(T* lhs, T rhs, unsigned int N){
//calculate the 1D index for this thread
int i = blockIdx.x * blockDim.x + threadIdx.x;
if(i < N)
lhs[i] *= rhs;
}
template<typename T>
void gpu_multiply(T* lhs, T rhs, unsigned int N){
//get the maximum number of threads per block for the CUDA device
int threads = stim::maxThreadsPerBlock();
//calculate the number of blocks
int blocks = N / threads + (N%threads == 0 ? 0:1);
//call the kernel to do the multiplication
cuda_multiply <<< blocks, threads >>>(lhs, rhs, N);
}
template<typename T>
void cpu_multiply(T* lhs, T rhs, unsigned int N){
//calculate the number of bytes in the array
unsigned int bytes = N * sizeof(T);
//allocate memory on the GPU for the array
T* gpuLHS;
HANDLE_ERROR( cudaMalloc(&gpuLHS, bytes) );
//copy the array to the GPU
HANDLE_ERROR( cudaMemcpy(gpuLHS, lhs, bytes, cudaMemcpyHostToDevice) );
//call the GPU version of this function
gpu_multiply<T>(gpuLHS, rhs, N);
//copy the array back to the CPU
HANDLE_ERROR( cudaMemcpy(lhs, gpuLHS, bytes, cudaMemcpyDeviceToHost) );
//free allocated memory
cudaFree(gpuLHS);
}
// array .* array multiplication
template<typename T>
__global__ void cuda_multiply(T* ptr1, T* ptr2, T* product, unsigned int N){
//calculate the 1D index for this thread
int idx = blockIdx.x * blockDim.x + threadIdx.x;
if(idx < N){
product[idx] = ptr1[idx] * ptr2[idx];
}
}
template<typename T>
void gpu_multiply(T* ptr1, T* ptr2, T* product, unsigned int N){
//get the maximum number of threads per block for the CUDA device
int threads = stim::maxThreadsPerBlock();
//calculate the number of blocks
int blocks = N / threads + 1;
//call the kernel to do the multiplication
cuda_multiply <<< blocks, threads >>>(ptr1, ptr2, product, N);
}
template<typename T>
void cpu_multiply(T* ptr1, T* ptr2, T* cpu_product, unsigned int N){
//allocate memory on the GPU for the array
T* gpu_ptr1;
T* gpu_ptr2;
T* gpu_product;
HANDLE_ERROR( cudaMalloc( &gpu_ptr1, N * sizeof(T) ) );
HANDLE_ERROR( cudaMalloc( &gpu_ptr2, N * sizeof(T) ) );
HANDLE_ERROR( cudaMalloc( &gpu_product, N * sizeof(T) ) );
//copy the array to the GPU
HANDLE_ERROR( cudaMemcpy( gpu_ptr1, ptr1, N * sizeof(T), cudaMemcpyHostToDevice) );
HANDLE_ERROR( cudaMemcpy( gpu_ptr2, ptr2, N * sizeof(T), cudaMemcpyHostToDevice) );
//call the GPU version of this function
gpu_multiply<T>(gpu_ptr1, gpu_ptr2 ,gpu_product, N);
//copy the array back to the CPU
HANDLE_ERROR( cudaMemcpy( cpu_product, gpu_product, N * sizeof(T), cudaMemcpyDeviceToHost) );
//free allocated memory
cudaFree(gpu_ptr1);
cudaFree(gpu_ptr2);
cudaFree(gpu_product);
}
}
}
#endif