-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathsacCudaCalls.cpp
199 lines (176 loc) · 6.39 KB
/
sacCudaCalls.cpp
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
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
//
#include "uebDAsrPFuebpgdecls.h"
//#include "uebDAdafunctions.h"
//#include <time.h>
//#include <queue>
#pragma warning(disable : 4996)
void cuda_checkERR(cudaError_t err);
void checkDeviceMemory();
void estimateThroughput(size_t dataSize, clock_t beginTime, clock_t endTime);
__host__ __device__
void SACGridDA::runSACGrid_Ens(int grdIndx, int nEns, float *pxvEns)
{
float edmndOrg = edmnd;
for (int ie = 0; ie < nEns; ie++)
{
edmnd = edmndOrg;
fland1Device(pxvEns[grdIndx * nEns + ie], dtday, aesc, sacstEns[ie], sacpar,
edmnd, surfEns[ie], grndEns[ie], tet,
sif, bfp, bfs, ssur, sdro, sperc, hrapx, hrapy, error);
}
}
__global__
void do_sac_EnsDevice_Ens(int npix, int nEns, SACGridDA *sacGridArray, float *pxvEns)
{
int i = blockIdx.x*blockDim.x + threadIdx.x;
if (i < npix)
{
sacGridArray[i].runSACGrid_Ens(i, nEns, pxvEns);
} /* End of grid loop */
} /* End of program */
int do_sac_EnsHost_Ens(int threadsPerBlock, int npix, int nEns, SACGridDA *sacGridArray, float *pxvEns)
{
//std::cout << " here in host call_8 " << std::endl;
int curDv, curStr;
//memory on device
//checkDeviceMemory()
cudaError_t err = cudaSuccess;
err = cudaGetDevice(&curDv);
cuda_checkERR(err);
//std::cout << " current device = " << curDv << std::endl;
cudaStream_t oStream;
//cudaSetDevice(cudIndx);
err = cudaStreamCreate(&oStream);
cuda_checkERR(err);
//std::cout << " current stream = " << oStream << std::endl;
//memory on device
//checkDeviceMemory();
SACGridDA *dev_sacGridArray = NULL;
err = cudaMalloc(&dev_sacGridArray, npix*sizeof(SACGridDA));
cuda_checkERR(err);
//std::cout << " device memory alloc ";
err = cudaMemcpyAsync(dev_sacGridArray, sacGridArray, npix*sizeof(SACGridDA), cudaMemcpyHostToDevice, oStream); // != cudaSuccess)
cuda_checkERR(err);
err = cudaStreamSynchronize(oStream);
cuda_checkERR(err);
//
float *dev_pxvEns = NULL;
err = cudaMalloc(&dev_pxvEns, nEns*npix*sizeof(float));
cuda_checkERR(err);
err = cudaMemcpyAsync(dev_pxvEns, pxvEns, nEns*npix*sizeof(float), cudaMemcpyHostToDevice, oStream); // != cudaSuccess)
cuda_checkERR(err);
err = cudaStreamSynchronize(oStream);
cuda_checkERR(err);
//std::cout << " device data copy ";
//dataCopy_Time += (MPI::Wtime() - intermStart_Time);
//intermStart_Time = MPI::Wtime();
//memory on device
//checkDeviceMemory();
// Launch Kernel
int blocksPerGrid = (npix + threadsPerBlock - 1) / threadsPerBlock;
do_sac_EnsDevice_Ens << < blocksPerGrid, threadsPerBlock, 0 >> > (npix, nEns, dev_sacGridArray, dev_pxvEns);
//synchronization
err = cudaStreamSynchronize(oStream);
cuda_checkERR(err);
//computeRun_Time += (MPI::Wtime() - intermStart_Time);
//intermStart_Time = MPI::Wtime();
//std::cout << " finished device compute tasks " << std::endl;
//copy data back
err = cudaMemcpyAsync(sacGridArray, dev_sacGridArray, npix*sizeof(SACGridDA), cudaMemcpyDeviceToHost, oStream);// != cudaSuccess)
cuda_checkERR(err);
err = cudaStreamSynchronize(oStream);// cudaDeviceSynchronize();
cuda_checkERR(err);
//dataCopy_Time += (MPI::Wtime() - intermStart_Time);
//intermStart_Time = MPI::Wtime();
//std::cout << " data copied to host "; //cout << err << endl;
//checkDeviceMemory();
//free device memory
err = cudaFree(dev_sacGridArray); // npsac_def*npix*sizeof(float));
cuda_checkERR(err);
err = cudaFree(dev_pxvEns); // npsac_def*npix*sizeof(float));
cuda_checkERR(err);
//std::cout <<" Device memory freed ";
//memory on device
//checkDeviceMemory();
//clear stream
err = cudaStreamDestroy(oStream);
cuda_checkERR(err);
return 0;
}
//
__global__
void do_sac_EnsDevice(int npix, SACGridDA *sacGridArray)
{
int i = blockIdx.x*blockDim.x + threadIdx.x;
if (i < npix)
{
sacGridArray[i].runSACGrid();
} /* End of grid loop */
} /* End of program */
int do_sac_EnsHost(int threadsPerBlock, int npix, SACGridDA *sacGridArray)
{
//std::cout << " here in host call_8 " << std::endl;
int curDv, curStr;
//memory on device
//checkDeviceMemory()
cudaError_t err = cudaSuccess;
err = cudaGetDevice(&curDv);
cuda_checkERR(err);
//std::cout << " current device = " << curDv << std::endl;
cudaStream_t oStream;
//cudaSetDevice(cudIndx);
err = cudaStreamCreate(&oStream);
cuda_checkERR(err);
//std::cout << " current stream = " << oStream << std::endl;
//memory on device
//checkDeviceMemory();
SACGridDA *dev_sacGridArray = NULL;
err = cudaMalloc(&dev_sacGridArray, npix*sizeof(SACGridDA));
cuda_checkERR(err);
//std::cout << " device memory alloc ";
err = cudaMemcpyAsync(dev_sacGridArray, sacGridArray, npix*sizeof(SACGridDA), cudaMemcpyHostToDevice, oStream); // != cudaSuccess)
cuda_checkERR(err);
err = cudaStreamSynchronize(oStream);
cuda_checkERR(err);
//std::cout << " device data copy ";
//dataCopy_Time += (MPI::Wtime() - intermStart_Time);
//intermStart_Time = MPI::Wtime();
//memory on device
//checkDeviceMemory();
// Launch Kernel
int blocksPerGrid = (npix + threadsPerBlock - 1) / threadsPerBlock;
do_sac_EnsDevice << < blocksPerGrid, threadsPerBlock, 0 >> > (npix, dev_sacGridArray);
//synchronization
err = cudaStreamSynchronize(oStream);
cuda_checkERR(err);
//computeRun_Time += (MPI::Wtime() - intermStart_Time);
//intermStart_Time = MPI::Wtime();
//std::cout << " finished device compute tasks " << std::endl;
//copy data back
err = cudaMemcpyAsync(sacGridArray, dev_sacGridArray, npix*sizeof(SACGridDA), cudaMemcpyDeviceToHost, oStream);// != cudaSuccess)
cuda_checkERR(err);
err = cudaStreamSynchronize(oStream);// cudaDeviceSynchronize();
cuda_checkERR(err);
//dataCopy_Time += (MPI::Wtime() - intermStart_Time);
//intermStart_Time = MPI::Wtime();
//std::cout << " data copied to host "; //cout << err << endl;
//checkDeviceMemory();
//free device memory
err = cudaFree(dev_sacGridArray); // npsac_def*npix*sizeof(float));
cuda_checkERR(err);
//std::cout <<" Device memory freed ";
//memory on device
//checkDeviceMemory();
//clear stream
err = cudaStreamDestroy(oStream);
cuda_checkERR(err);
return 0;
}
void do_sac_EnsHost_PC(int npix, SACGridDA *sacGridArray)
{
for (int i = 0; i<npix; i++)
//if (i < npix)
{
sacGridArray[i].runSACGrid();
} /* End of grid loop */
} /* End of program */