| 1 |
/* Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. |
| 2 |
* |
| 3 |
* Redistribution and use in source and binary forms, with or without |
| 4 |
* modification, are permitted provided that the following conditions |
| 5 |
* are met: |
| 6 |
* * Redistributions of source code must retain the above copyright |
| 7 |
* notice, this list of conditions and the following disclaimer. |
| 8 |
* * Redistributions in binary form must reproduce the above copyright |
| 9 |
* notice, this list of conditions and the following disclaimer in the |
| 10 |
* documentation and/or other materials provided with the distribution. |
| 11 |
* * Neither the name of NVIDIA CORPORATION nor the names of its |
| 12 |
* contributors may be used to endorse or promote products derived |
| 13 |
* from this software without specific prior written permission. |
| 14 |
* |
| 15 |
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY |
| 16 |
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
| 17 |
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR |
| 18 |
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR |
| 19 |
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, |
| 20 |
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, |
| 21 |
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR |
| 22 |
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY |
| 23 |
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT |
| 24 |
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE |
| 25 |
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
| 26 |
*/ |
| 27 |
|
| 28 |
/////////////////////////////////////////////////////////////////////////////// |
| 29 |
// Monte Carlo: Estimate Pi |
| 30 |
// ======================== |
| 31 |
// |
| 32 |
// This sample demonstrates a very simple Monte Carlo estimation for Pi. |
| 33 |
// |
| 34 |
// This file, main.cpp, contains the setup information to run the test, for |
| 35 |
// example parsing the command line and integrating this sample with the |
| 36 |
// samples framework. As such it is perhaps less interesting than the guts of |
| 37 |
// the sample. Readers wishing to skip the clutter are advised to skip straight |
| 38 |
// to Test.operator() in test.cpp. |
| 39 |
/////////////////////////////////////////////////////////////////////////////// |
| 40 |
|
| 41 |
#include <cuda_runtime.h> |
| 42 |
#include <helper_cuda.h> |
| 43 |
|
| 44 |
#include <iomanip> |
| 45 |
#include <iostream> |
| 46 |
#include <stdexcept> |
| 47 |
|
| 48 |
#include "../inc/test.h" |
| 49 |
|
| 50 |
// SDK information |
| 51 |
static const char *printfFile = "MonteCarloEstimatePiInlineP.txt"; |
| 52 |
|
| 53 |
// Forward declarations |
| 54 |
void showHelp(const int argc, const char **argv); |
| 55 |
template <typename Real> |
| 56 |
void runTest(int argc, const char **argv); |
| 57 |
|
| 58 |
int main(int argc, char **argv) { |
| 59 |
using std::invalid_argument; |
| 60 |
using std::string; |
| 61 |
|
| 62 |
// Open the log file |
| 63 |
printf("Monte Carlo Estimate Pi (with inline PRNG)\n"); |
| 64 |
printf("==========================================\n\n"); |
| 65 |
|
| 66 |
// If help flag is set, display help and exit immediately |
| 67 |
if (checkCmdLineFlag(argc, (const char **)argv, "help")) { |
| 68 |
printf("Displaying help on console\n"); |
| 69 |
showHelp(argc, (const char **)argv); |
| 70 |
exit(EXIT_SUCCESS); |
| 71 |
} |
| 72 |
|
| 73 |
// Check the precision (checked against the device capability later) |
| 74 |
try { |
| 75 |
char *value; |
| 76 |
|
| 77 |
if (getCmdLineArgumentString(argc, (const char **)argv, "precision", |
| 78 |
&value)) { |
| 79 |
// Check requested precision is valid |
| 80 |
string prec(value); |
| 81 |
|
| 82 |
if (prec.compare("single") == 0 || prec.compare("\"single\"") == 0) { |
| 83 |
runTest<float>(argc, (const char **)argv); |
| 84 |
} else if (prec.compare("double") == 0 || |
| 85 |
prec.compare("\"double\"") == 0) { |
| 86 |
runTest<double>(argc, (const char **)argv); |
| 87 |
} else { |
| 88 |
printf( |
| 89 |
"specified precision (%s) is invalid, must be \"single\" or " |
| 90 |
"\"double\".\n", |
| 91 |
value); |
| 92 |
throw invalid_argument("precision"); |
| 93 |
} |
| 94 |
} else { |
| 95 |
runTest<float>(argc, (const char **)argv); |
| 96 |
} |
| 97 |
} catch (invalid_argument &e) { |
| 98 |
printf("invalid command line argument (%s)\n", e.what()); |
| 99 |
exit(EXIT_FAILURE); |
| 100 |
} |
| 101 |
|
| 102 |
// Finish |
| 103 |
exit(EXIT_SUCCESS); |
| 104 |
} |
| 105 |
|
| 106 |
template <typename Real> |
| 107 |
void runTest(int argc, const char **argv) { |
| 108 |
using std::invalid_argument; |
| 109 |
using std::runtime_error; |
| 110 |
|
| 111 |
try { |
| 112 |
Test<Real> test; |
| 113 |
int deviceCount = 0; |
| 114 |
cudaError_t cudaResult = cudaSuccess; |
| 115 |
|
| 116 |
// by default specify GPU Device == 0 |
| 117 |
test.device = 0; |
| 118 |
|
| 119 |
// Get number of available devices |
| 120 |
cudaResult = cudaGetDeviceCount(&deviceCount); |
| 121 |
|
| 122 |
if (cudaResult != cudaSuccess) { |
| 123 |
printf("could not get device count (%s).\n", |
| 124 |
cudaGetErrorString(cudaResult)); |
| 125 |
throw runtime_error("cudaGetDeviceCount"); |
| 126 |
} |
| 127 |
|
| 128 |
// (default parameters) |
| 129 |
test.numSims = k_sims_qa; |
| 130 |
test.threadBlockSize = k_bsize_qa; |
| 131 |
test.seed = k_seed_def; |
| 132 |
|
| 133 |
{ |
| 134 |
char *value = 0; |
| 135 |
|
| 136 |
if (getCmdLineArgumentString(argc, argv, "device", &value)) { |
| 137 |
test.device = (int)atoi(value); |
| 138 |
|
| 139 |
if (test.device >= deviceCount) { |
| 140 |
printf( |
| 141 |
"invalid target device specified on command line (device %d does " |
| 142 |
"not exist).\n", |
| 143 |
test.device); |
| 144 |
throw invalid_argument("device"); |
| 145 |
} |
| 146 |
} else { |
| 147 |
test.device = gpuGetMaxGflopsDeviceId(); |
| 148 |
} |
| 149 |
|
| 150 |
if (getCmdLineArgumentString(argc, argv, "sims", &value)) { |
| 151 |
test.numSims = (unsigned int)atoi(value); |
| 152 |
printf("number of simulations = %d\n", test.numSims); |
| 153 |
|
| 154 |
if (test.numSims < k_sims_min || test.numSims > k_sims_max) { |
| 155 |
printf( |
| 156 |
"specified number of simulations (%d) is invalid, must be " |
| 157 |
"between %d and %d.\n", |
| 158 |
test.numSims, k_sims_min, k_sims_max); |
| 159 |
throw invalid_argument("sims"); |
| 160 |
} |
| 161 |
} else { |
| 162 |
test.numSims = k_sims_def; |
| 163 |
} |
| 164 |
|
| 165 |
if (getCmdLineArgumentString(argc, argv, "block-size", &value)) { |
| 166 |
// Determine max threads per block |
| 167 |
cudaDeviceProp deviceProperties; |
| 168 |
cudaResult = cudaGetDeviceProperties(&deviceProperties, test.device); |
| 169 |
|
| 170 |
if (cudaResult != cudaSuccess) { |
| 171 |
printf("unable to get device properties for device %d.\n", |
| 172 |
test.device); |
| 173 |
throw runtime_error("cudaGetDeviceProperties"); |
| 174 |
} |
| 175 |
|
| 176 |
// Check requested size is valid |
| 177 |
test.threadBlockSize = (unsigned int)atoi(value); |
| 178 |
printf("block size = %d\n", test.threadBlockSize); |
| 179 |
|
| 180 |
if (test.threadBlockSize < k_bsize_min || |
| 181 |
test.threadBlockSize > static_cast<unsigned int>( |
| 182 |
deviceProperties.maxThreadsPerBlock)) { |
| 183 |
printf( |
| 184 |
"specified block size (%d) is invalid, must be between %d and %d " |
| 185 |
"for device %d.\n", |
| 186 |
test.threadBlockSize, k_bsize_min, |
| 187 |
deviceProperties.maxThreadsPerBlock, test.device); |
| 188 |
throw invalid_argument("block-size"); |
| 189 |
} |
| 190 |
|
| 191 |
if (test.threadBlockSize & test.threadBlockSize - 1) { |
| 192 |
printf( |
| 193 |
"specified block size (%d) is invalid, must be a power of two " |
| 194 |
"(see reduction function).\n", |
| 195 |
test.threadBlockSize); |
| 196 |
throw invalid_argument("block-size"); |
| 197 |
} |
| 198 |
} else { |
| 199 |
test.threadBlockSize = k_bsize_def; |
| 200 |
} |
| 201 |
|
| 202 |
if (getCmdLineArgumentString(argc, argv, "seed", &value)) { |
| 203 |
// Check requested seed is valid |
| 204 |
test.seed = (unsigned int)atoi(value); |
| 205 |
printf("seed = %d\n", test.seed); |
| 206 |
|
| 207 |
if (test.seed == 0) { |
| 208 |
printf("specified seed (%d) is invalid, must be non-zero.\n", |
| 209 |
test.seed); |
| 210 |
throw invalid_argument("seed"); |
| 211 |
} |
| 212 |
} else { |
| 213 |
test.seed = k_seed_def; |
| 214 |
} |
| 215 |
} |
| 216 |
|
| 217 |
// Execute |
| 218 |
test(); |
| 219 |
} catch (invalid_argument &e) { |
| 220 |
printf("invalid command line argument (%s)\n", e.what()); |
| 221 |
exit(EXIT_FAILURE); |
| 222 |
} catch (runtime_error &e) { |
| 223 |
printf("runtime error (%s)\n", e.what()); |
| 224 |
exit(EXIT_FAILURE); |
| 225 |
} |
| 226 |
} |
| 227 |
|
| 228 |
void showHelp(int argc, const char **argv) { |
| 229 |
using std::cout; |
| 230 |
using std::endl; |
| 231 |
using std::left; |
| 232 |
using std::setw; |
| 233 |
|
| 234 |
if (argc > 0) { |
| 235 |
cout << endl << argv[0] << endl; |
| 236 |
} |
| 237 |
|
| 238 |
cout << endl << "Syntax:" << endl; |
| 239 |
cout << left; |
| 240 |
cout << " " << setw(20) << "--device=<device>" |
| 241 |
<< "Specify device to use for execution" << endl; |
| 242 |
cout << " " << setw(20) << "--sims=<N>" |
| 243 |
<< "Specify number of Monte Carlo simulations" << endl; |
| 244 |
cout << " " << setw(20) << "--block-size=<N>" |
| 245 |
<< "Specify number of threads per block" << endl; |
| 246 |
cout << " " << setw(20) << "--seed=<N>" |
| 247 |
<< "Specify the seed to use for the random number generator" << endl; |
| 248 |
cout << " " << setw(20) << "--precision=<P>" |
| 249 |
<< "Specify the precision (\"single\" or \"double\")" << endl; |
| 250 |
cout << endl; |
| 251 |
cout << " " << setw(20) << "--noprompt" |
| 252 |
<< "Skip prompt before exit" << endl; |
| 253 |
cout << endl; |
| 254 |
} |