018b00d6
 David Mayerich
adapted bsds500 f...
|
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
|
#ifndef STIM_CUDA_ARRAY_ATAN2_H
#define STIM_CUDA_ARRAY_ATAN2_H
#include <iostream>
#include <cuda.h>
#include <cmath>
#include <stim/cuda/cudatools.h>
namespace stim{
namespace cuda{
template<typename T>
__global__ void cuda_atan2(T* y, T* x, T* r, unsigned int N){
//calculate the 1D index for this thread
int idx = blockIdx.x * blockDim.x + threadIdx.x;
if(idx < N){
r[idx] = atan2(y[idx], x[idx]);
}
}
template<typename T>
void gpu_atan2(T* y, T* x, T* r, 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_atan2 <<< blocks, threads >>>(y, x, r, N);
}
template<typename T>
void cpu_atan2(T* y, T* x, T* cpu_r, unsigned int N){
//allocate memory on the GPU for the array
T* gpu_x;
T* gpu_y;
T* gpu_r;
HANDLE_ERROR( cudaMalloc( &gpu_x, N * sizeof(T) ) );
HANDLE_ERROR( cudaMalloc( &gpu_y, N * sizeof(T) ) );
HANDLE_ERROR( cudaMalloc( &gpu_r, N * sizeof(T) ) );
//copy the array to the GPU
HANDLE_ERROR( cudaMemcpy( gpu_x, x, N * sizeof(T), cudaMemcpyHostToDevice) );
HANDLE_ERROR( cudaMemcpy( gpu_y, y, N * sizeof(T), cudaMemcpyHostToDevice) );
//call the GPU version of this function
gpu_atan2<T>(gpu_y, gpu_x ,gpu_r, N);
//copy the array back to the CPU
HANDLE_ERROR( cudaMemcpy( cpu_r, gpu_r, N * sizeof(T), cudaMemcpyDeviceToHost) );
//free allocated memory
cudaFree(gpu_x);
cudaFree(gpu_y);
cudaFree(gpu_r);
}
}
}
#endif
|
