fiber/examples/cuda/multiple_streams.cu
2017-11-03 07:59:26 +01:00

113 lines
5.0 KiB
Plaintext

// Copyright Oliver Kowalke 2013.
// Distributed under the Boost Software License, Version 1.0.
// (See accompanying file LICENSE_1_0.txt or copy at
// http://www.boost.org/LICENSE_1_0.txt)
#include <chrono>
#include <cstdlib>
#include <iostream>
#include <memory>
#include <random>
#include <tuple>
#include <cuda.h>
#include <boost/assert.hpp>
#include <boost/bind.hpp>
#include <boost/intrusive_ptr.hpp>
#include <boost/fiber/all.hpp>
#include <boost/fiber/cuda/waitfor.hpp>
__global__
void vector_add( int * a, int * b, int * c, int size) {
int idx = threadIdx.x + blockIdx.x * blockDim.x;
if ( idx < size) {
c[idx] = a[idx] + b[idx];
}
}
int main() {
try {
bool done = false;
boost::fibers::fiber f1( [&done]{
std::cout << "f1: entered" << std::endl;
try {
cudaStream_t stream0, stream1;
cudaStreamCreate( & stream0);
cudaStreamCreate( & stream1);
int size = 1024 * 1024;
int full_size = 20 * size;
int * host_a, * host_b, * host_c;
cudaHostAlloc( & host_a, full_size * sizeof( int), cudaHostAllocDefault);
cudaHostAlloc( & host_b, full_size * sizeof( int), cudaHostAllocDefault);
cudaHostAlloc( & host_c, full_size * sizeof( int), cudaHostAllocDefault);
int * dev_a0, * dev_b0, * dev_c0;
int * dev_a1, * dev_b1, * dev_c1;
cudaMalloc( & dev_a0, size * sizeof( int) );
cudaMalloc( & dev_b0, size * sizeof( int) );
cudaMalloc( & dev_c0, size * sizeof( int) );
cudaMalloc( & dev_a1, size * sizeof( int) );
cudaMalloc( & dev_b1, size * sizeof( int) );
cudaMalloc( & dev_c1, size * sizeof( int) );
std::minstd_rand generator;
std::uniform_int_distribution<> distribution(1, 6);
for ( int i = 0; i < full_size; ++i) {
host_a[i] = distribution( generator);
host_b[i] = distribution( generator);
}
for ( int i = 0; i < full_size; i += 2 * size) {
cudaMemcpyAsync( dev_a0, host_a + i, size * sizeof( int), cudaMemcpyHostToDevice, stream0);
cudaMemcpyAsync( dev_a1, host_a + i + size, size * sizeof( int), cudaMemcpyHostToDevice, stream1);
cudaMemcpyAsync( dev_b0, host_b + i, size * sizeof( int), cudaMemcpyHostToDevice, stream0);
cudaMemcpyAsync( dev_b1, host_b + i + size, size * sizeof( int), cudaMemcpyHostToDevice, stream1);
vector_add<<< size / 256, 256, 0, stream0 >>>( dev_a0, dev_b0, dev_c0, size);
vector_add<<< size / 256, 256, 0, stream1 >>>( dev_a1, dev_b1, dev_c1, size);
cudaMemcpyAsync( host_c + i, dev_c0, size * sizeof( int), cudaMemcpyDeviceToHost, stream0);
cudaMemcpyAsync( host_c + i + size, dev_c1, size * sizeof( int), cudaMemcpyDeviceToHost, stream1);
}
auto results = boost::fibers::cuda::waitfor_all( stream0, stream1);
for ( auto & result : results) {
BOOST_ASSERT( stream0 == std::get< 0 >( result) || stream1 == std::get< 0 >( result) );
BOOST_ASSERT( cudaSuccess == std::get< 1 >( result) );
}
std::cout << "f1: GPU computation finished" << std::endl;
cudaFreeHost( host_a);
cudaFreeHost( host_b);
cudaFreeHost( host_c);
cudaFree( dev_a0);
cudaFree( dev_b0);
cudaFree( dev_c0);
cudaFree( dev_a1);
cudaFree( dev_b1);
cudaFree( dev_c1);
cudaStreamDestroy( stream0);
cudaStreamDestroy( stream1);
done = true;
} catch ( std::exception const& ex) {
std::cerr << "exception: " << ex.what() << std::endl;
}
std::cout << "f1: leaving" << std::endl;
});
boost::fibers::fiber f2([&done]{
std::cout << "f2: entered" << std::endl;
while ( ! done) {
std::cout << "f2: sleeping" << std::endl;
boost::this_fiber::sleep_for( std::chrono::milliseconds( 1 ) );
}
std::cout << "f2: leaving" << std::endl;
});
f1.join();
f2.join();
std::cout << "done." << std::endl;
return EXIT_SUCCESS;
} catch ( std::exception const& e) {
std::cerr << "exception: " << e.what() << std::endl;
} catch (...) {
std::cerr << "unhandled exception" << std::endl;
}
return EXIT_FAILURE;
}