NASP/cuda-samples
85
0
Fork 0
mirror of https://github.com/NVIDIA/cuda-samples.git synced 2026-10-23 21:28:17 +08:00
Code Issues Packages Projects Releases Wiki Activity

add and update samples with CUDA 11.3 support

Browse Source
This commit is contained in:
Rutwik Choughule
2021-04-16 11:54:26 +05:30
parent 067cb65523
commit 568b39bd5b
214 changed files with 6590 additions and 3856 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
Samples/conjugateGradientMultiDeviceCG/README.md
View File

@@ -30,7 +30,7 @@ cudaMemAdvise, cudaMemPrefetchAsync, cudaLaunchCooperativeKernelMultiDevice, cud
## Prerequisites
Download and install the [CUDA Toolkit 11.2](https://developer.nvidia.com/cuda-downloads) for your corresponding platform.
Download and install the [CUDA Toolkit 11.3](https://developer.nvidia.com/cuda-downloads) for your corresponding platform.
Make sure the dependencies mentioned in [Dependencies]() section above are installed.
## Build and Run

138
Samples/conjugateGradientMultiDeviceCG/conjugateGradientMultiDeviceCG.cu
View File

@@ -223,8 +223,10 @@ __device__ void gpuDotProduct(float *vecA, float *vecB, int size,
cg::sync(cta);
if (tile32.meta_group_rank() == 0) {
temp_sum = tile32.thread_rank() < tile32.meta_group_size() ? tmp[tile32.thread_rank()] : 0.0;
temp_sum = cg::reduce(tile32, temp_sum, cg::plus<double>());
temp_sum = tile32.thread_rank() < tile32.meta_group_size()
? tmp[tile32.thread_rank()]
: 0.0;
temp_sum = cg::reduce(tile32, temp_sum, cg::plus<double>());
if (tile32.thread_rank() == 0) {
atomicAdd(&grid_dot_result, temp_sum);
@@ -239,8 +241,9 @@ __device__ void gpuCopyVector(float *srcA, float *destB, int size,
}
}
__device__ void gpuScaleVectorAndSaxpy(float *x, float *y, float a, float scale, int size,
const cg::multi_grid_group &multi_grid) {
__device__ void gpuScaleVectorAndSaxpy(float *x, float *y, float a, float scale,
int size,
const cg::multi_grid_group &multi_grid) {
for (int i = multi_grid.thread_rank(); i < size; i += multi_grid.size()) {
y[i] = a * x[i] + scale * y[i];
}
@@ -360,10 +363,11 @@ std::multimap<std::pair<int, int>, int> getIdenticalGPUs() {
// Filter unsupported devices
if (deviceProp.cooperativeMultiDeviceLaunch &&
deviceProp.concurrentManagedAccess) {
identicalGpus.emplace(std::make_pair(deviceProp.major, deviceProp.minor), i);
identicalGpus.emplace(std::make_pair(deviceProp.major, deviceProp.minor),
i);
}
printf("GPU Device %d: \"%s\" with compute capability %d.%d\n", i,
deviceProp.name, deviceProp.major, deviceProp.minor);
deviceProp.name, deviceProp.major, deviceProp.minor);
}
return identicalGpus;
@@ -387,15 +391,17 @@ int main(int argc, char **argv) {
auto bestFit = std::make_pair(it, it);
// use std::distance to find the largest number of GPUs amongst architectures
auto distance = [](decltype(bestFit) p){return std::distance(p.first, p.second);};
auto distance = [](decltype(bestFit) p) {
return std::distance(p.first, p.second);
};
// Read each unique key/pair element in order
for (; it != end; it = gpusByArch.upper_bound(it->first)) {
// first and second are iterators bounded within the architecture group
auto testFit = gpusByArch.equal_range(it->first);
// Always use devices with highest architecture version or whichever has the most devices available
if (distance(bestFit) <= distance(testFit))
bestFit = testFit;
// Always use devices with highest architecture version or whichever has the
// most devices available
if (distance(bestFit) <= distance(testFit)) bestFit = testFit;
}
if (distance(bestFit) < kNumGpusRequired) {
@@ -408,33 +414,35 @@ int main(int argc, char **argv) {
std::set<int> bestFitDeviceIds;
// check & select peer-to-peer access capable GPU devices as enabling p2p access between participating
// check & select peer-to-peer access capable GPU devices as enabling p2p
// access between participating
// GPUs gives better performance for multi_grid sync.
for (auto itr = bestFit.first; itr != bestFit.second; itr++) {
int deviceId = itr->second;
checkCudaErrors(cudaSetDevice(deviceId));
std::for_each(itr, bestFit.second, [&deviceId, &bestFitDeviceIds](decltype(*itr) mapPair) {
if (deviceId != mapPair.second)
{
std::for_each(itr, bestFit.second, [&deviceId, &bestFitDeviceIds,
&kNumGpusRequired](
decltype(*itr) mapPair) {
if (deviceId != mapPair.second) {
int access = 0;
checkCudaErrors(cudaDeviceCanAccessPeer(&access, deviceId, mapPair.second));
printf("Device=%d %s Access Peer Device=%d\n", deviceId, access ? "CAN" : "CANNOT", mapPair.second);
checkCudaErrors(
cudaDeviceCanAccessPeer(&access, deviceId, mapPair.second));
printf("Device=%d %s Access Peer Device=%d\n", deviceId,
access ? "CAN" : "CANNOT", mapPair.second);
if (access && bestFitDeviceIds.size() < kNumGpusRequired) {
bestFitDeviceIds.emplace(deviceId);
bestFitDeviceIds.emplace(mapPair.second);
}
else {
} else {
printf("Ignoring device %i (max devices exceeded)\n", mapPair.second);
}
}
});
if (bestFitDeviceIds.size() >= kNumGpusRequired)
{
if (bestFitDeviceIds.size() >= kNumGpusRequired) {
printf("Selected p2p capable devices - ");
for (auto devicesItr = bestFitDeviceIds.begin(); devicesItr != bestFitDeviceIds.end(); devicesItr++)
{
for (auto devicesItr = bestFitDeviceIds.begin();
devicesItr != bestFitDeviceIds.end(); devicesItr++) {
printf("deviceId = %d ", *devicesItr);
}
printf("\n");
@@ -442,33 +450,34 @@ int main(int argc, char **argv) {
}
}
// if bestFitDeviceIds.size() == 0 it means the GPUs in system are not p2p capable,
// if bestFitDeviceIds.size() == 0 it means the GPUs in system are not p2p
// capable,
// hence we add it without p2p capability check.
if (!bestFitDeviceIds.size())
{
printf("Devices involved are not p2p capable.. selecting %zu of them\n", kNumGpusRequired);
std::for_each(bestFit.first, bestFit.second, [&bestFitDeviceIds](decltype(*bestFit.first) mapPair) {
if (bestFitDeviceIds.size() < kNumGpusRequired) {
bestFitDeviceIds.emplace(mapPair.second);
}
else {
printf("Ignoring device %i (max devices exceeded)\n", mapPair.second);
}
// Insert the sequence into the deviceIds set
});
}
else
{
// perform cudaDeviceEnablePeerAccess in both directions for all participating devices
// of a cudaLaunchCooperativeKernelMultiDevice call this gives better performance for multi_grid sync.
for (auto p1_itr = bestFitDeviceIds.begin(); p1_itr != bestFitDeviceIds.end(); p1_itr++)
{
if (!bestFitDeviceIds.size()) {
printf("Devices involved are not p2p capable.. selecting %zu of them\n",
kNumGpusRequired);
std::for_each(bestFit.first, bestFit.second,
[&bestFitDeviceIds,
&kNumGpusRequired](decltype(*bestFit.first) mapPair) {
if (bestFitDeviceIds.size() < kNumGpusRequired) {
bestFitDeviceIds.emplace(mapPair.second);
} else {
printf("Ignoring device %i (max devices exceeded)\n",
mapPair.second);
}
// Insert the sequence into the deviceIds set
});
} else {
// perform cudaDeviceEnablePeerAccess in both directions for all
// participating devices of a cudaLaunchCooperativeKernelMultiDevice call
// this gives better performance for multi_grid sync.
for (auto p1_itr = bestFitDeviceIds.begin();
p1_itr != bestFitDeviceIds.end(); p1_itr++) {
checkCudaErrors(cudaSetDevice(*p1_itr));
for (auto p2_itr = bestFitDeviceIds.begin(); p2_itr != bestFitDeviceIds.end(); p2_itr++)
{
if (*p1_itr != *p2_itr)
{
checkCudaErrors(cudaDeviceEnablePeerAccess(*p2_itr, 0 ));
for (auto p2_itr = bestFitDeviceIds.begin();
p2_itr != bestFitDeviceIds.end(); p2_itr++) {
if (*p1_itr != *p2_itr) {
checkCudaErrors(cudaDeviceEnablePeerAccess(*p2_itr, 0));
checkCudaErrors(cudaSetDevice(*p1_itr));
}
}
@@ -518,7 +527,7 @@ int main(int argc, char **argv) {
std::cout << "\nRunning on GPUs = " << kNumGpusRequired << std::endl;
cudaStream_t nStreams[kNumGpusRequired];
int sMemSize = sizeof(double) * ((THREADS_PER_BLOCK/32) + 1);
int sMemSize = sizeof(double) * ((THREADS_PER_BLOCK / 32) + 1);
int numBlocksPerSm = INT_MAX;
int numThreads = THREADS_PER_BLOCK;
int numSms = INT_MAX;
@@ -530,17 +539,16 @@ int main(int argc, char **argv) {
checkCudaErrors(cudaSetDevice(*deviceId));
checkCudaErrors(cudaGetDeviceProperties(&deviceProp, *deviceId));
int numBlocksPerSm_current=0;
int numBlocksPerSm_current = 0;
checkCudaErrors(cudaOccupancyMaxActiveBlocksPerMultiprocessor(
&numBlocksPerSm_current, multiGpuConjugateGradient, numThreads, sMemSize));
&numBlocksPerSm_current, multiGpuConjugateGradient, numThreads,
sMemSize));
if (numBlocksPerSm > numBlocksPerSm_current)
{
numBlocksPerSm = numBlocksPerSm_current;
if (numBlocksPerSm > numBlocksPerSm_current) {
numBlocksPerSm = numBlocksPerSm_current;
}
if (numSms > deviceProp.multiProcessorCount)
{
numSms = deviceProp.multiProcessorCount;
if (numSms > deviceProp.multiProcessorCount) {
numSms = deviceProp.multiProcessorCount;
}
deviceId++;
}
@@ -554,7 +562,7 @@ int main(int argc, char **argv) {
int device_count = 0;
int totalThreadsPerGPU = numSms * numBlocksPerSm * THREADS_PER_BLOCK;
deviceId = bestFitDeviceIds.begin();;
deviceId = bestFitDeviceIds.begin();
while (deviceId != bestFitDeviceIds.end()) {
checkCudaErrors(cudaSetDevice(*deviceId));
checkCudaErrors(cudaStreamCreate(&nStreams[device_count]));
@@ -621,14 +629,15 @@ int main(int argc, char **argv) {
printf("Total threads per GPU = %d numBlocksPerSm = %d\n",
numSms * numBlocksPerSm * THREADS_PER_BLOCK, numBlocksPerSm);
dim3 dimGrid(numSms * numBlocksPerSm, 1, 1), dimBlock(THREADS_PER_BLOCK, 1, 1);
dim3 dimGrid(numSms * numBlocksPerSm, 1, 1),
dimBlock(THREADS_PER_BLOCK, 1, 1);
void *kernelArgs[] = {
(void *)&I, (void *)&J, (void *)&val, (void *)&x,
(void *)&Ax, (void *)&p, (void *)&r, (void *)&dot_result,
(void *)&nz, (void *)&N, (void *)&tol,
};
cudaLaunchParams *launchParamsList = (cudaLaunchParams *)malloc(
sizeof(cudaLaunchParams) * kNumGpusRequired);
cudaLaunchParams *launchParamsList =
(cudaLaunchParams *)malloc(sizeof(cudaLaunchParams) * kNumGpusRequired);
for (int i = 0; i < kNumGpusRequired; i++) {
launchParamsList[i].func = (void *)multiGpuConjugateGradient;
launchParamsList[i].gridDim = dimGrid;
@@ -645,12 +654,11 @@ int main(int argc, char **argv) {
cudaCooperativeLaunchMultiDeviceNoPreSync |
cudaCooperativeLaunchMultiDeviceNoPostSync));
checkCudaErrors(cudaMemPrefetchAsync(x, sizeof(float) * N, cudaCpuDeviceId));
checkCudaErrors(
cudaMemPrefetchAsync(x, sizeof(float) * N, cudaCpuDeviceId));
checkCudaErrors(
cudaMemPrefetchAsync(dot_result, sizeof(double), cudaCpuDeviceId));
cudaMemPrefetchAsync(dot_result, sizeof(double), cudaCpuDeviceId));
deviceId = bestFitDeviceIds.begin();;
deviceId = bestFitDeviceIds.begin();
device_count = 0;
while (deviceId != bestFitDeviceIds.end()) {
checkCudaErrors(cudaSetDevice(*deviceId));
@@ -658,7 +666,7 @@ int main(int argc, char **argv) {
deviceId++;
}
r1 = *dot_result;
r1 = (float)*dot_result;
printf("GPU Final, residual = %e \n ", sqrt(r1));

4
Samples/conjugateGradientMultiDeviceCG/conjugateGradientMultiDeviceCG_vs2017.vcxproj
View File

@@ -38,7 +38,7 @@
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.props" />
<ImportGroup Label="ExtensionSettings">
<Import Project="$(CUDAPropsPath)\CUDA 11.2.props" />
<Import Project="$(CUDAPropsPath)\CUDA 11.3.props" />
</ImportGroup>
<ImportGroup Label="PropertySheets">
<Import Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" />
@@ -109,6 +109,6 @@
</ItemGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.targets" />
<ImportGroup Label="ExtensionTargets">
<Import Project="$(CUDAPropsPath)\CUDA 11.2.targets" />
<Import Project="$(CUDAPropsPath)\CUDA 11.3.targets" />
</ImportGroup>
</Project>

4
Samples/conjugateGradientMultiDeviceCG/conjugateGradientMultiDeviceCG_vs2019.vcxproj
View File

@@ -34,7 +34,7 @@
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.props" />
<ImportGroup Label="ExtensionSettings">
<Import Project="$(CUDAPropsPath)\CUDA 11.2.props" />
<Import Project="$(CUDAPropsPath)\CUDA 11.3.props" />
</ImportGroup>
<ImportGroup Label="PropertySheets">
<Import Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" />
@@ -105,6 +105,6 @@
</ItemGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.targets" />
<ImportGroup Label="ExtensionTargets">
<Import Project="$(CUDAPropsPath)\CUDA 11.2.targets" />
<Import Project="$(CUDAPropsPath)\CUDA 11.3.targets" />
</ImportGroup>
</Project>