2
0
mirror of https://github.com/boostorg/compute.git synced 2026-01-26 18:32:48 +00:00
Files
compute/perf/perf_saxpy.cpp
Kyle Lutz 07e4a6b3aa Remove BLAS functions
This removes the incomplete BLAS API functions.
2013-09-24 22:19:56 -04:00

96 lines
2.8 KiB
C++

//---------------------------------------------------------------------------//
// Copyright (c) 2013 Kyle Lutz <kyle.r.lutz@gmail.com>
//
// 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
//
// See http://kylelutz.github.com/compute for more information.
//---------------------------------------------------------------------------//
#include <algorithm>
#include <iostream>
#include <vector>
#include <boost/compute.hpp>
#include <boost/compute/detail/timer.hpp>
float rand_float()
{
return (float(rand()) / float(RAND_MAX)) * 1000.f;
}
// y <- alpha * x + y
void serial_saxpy(size_t n, float alpha, const float *x, float *y)
{
for(size_t i = 0; i < n; i++){
y[i] = alpha * x[i] + y[i];
}
}
int main(int argc, char *argv[])
{
using boost::compute::lambda::_1;
using boost::compute::lambda::_2;
size_t size = 1000;
if(argc >= 2){
size = boost::lexical_cast<size_t>(argv[1]);
}
std::cout << "size: " << size << std::endl;
float alpha = 2.5f;
// setup context and queue for the default device
boost::compute::device device = boost::compute::system::default_device();
boost::compute::context context(device);
boost::compute::command_queue queue(context, device);
// create vector of random numbers on the host
std::vector<float> host_x(size);
std::vector<float> host_y(size);
std::generate(host_x.begin(), host_x.end(), rand_float);
std::generate(host_y.begin(), host_y.end(), rand_float);
// create vector on the device and copy the data
boost::compute::vector<float> device_x(host_x.begin(), host_x.end(), context);
boost::compute::vector<float> device_y(host_y.begin(), host_y.end(), context);
boost::compute::detail::timer t;
boost::compute::transform(
device_x.begin(),
device_x.end(),
device_y.begin(),
device_y.begin(),
alpha * _1 + _2,
queue
);
queue.finish();
std::cout << "time: " << t.elapsed() << " ms" << std::endl;
// perform saxpy on host
serial_saxpy(size, alpha, &host_x[0], &host_y[0]);
// copy device_y to host_x
boost::compute::copy(device_y.begin(), device_y.end(), host_x.begin(), queue);
for(size_t i = 0; i < size; i++){
float host_value = host_y[i];
float device_value = host_x[i];
if(std::abs(device_value - host_value) > 1e-3){
std::cout << "ERROR: "
<< "value at " << i << " "
<< "device_value (" << device_value << ") "
<< "!= "
<< "host_value (" << host_value << ")"
<< std::endl;
return -1;
}
}
return 0;
}