You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
73 lines
2.1 KiB
73 lines
2.1 KiB
#include <iostream>
|
|
#include "../Common.h"
|
|
#include "Tests.h"
|
|
#include <random>
|
|
|
|
|
|
Square::Square() {
|
|
const float vmax = 1000.0f;
|
|
data_size = 4096;
|
|
std::random_device rd;
|
|
std::mt19937 gen(rd());
|
|
std::uniform_real_distribution<float> distr(0.0f, vmax);
|
|
|
|
for (size_t index = 0; index < data_size; ++index)
|
|
{
|
|
float val = distr(gen);
|
|
gpuResult.push_back(val);
|
|
sourceData.push_back(val);
|
|
}
|
|
}
|
|
void Square::gpu_compute(cl::Context* context, cl::CommandQueue* queue, cl::Program* program, cl::Event* Event) {
|
|
cl_int err = CL_SUCCESS;
|
|
|
|
// Get the kernel handle
|
|
cl::Kernel kernel(*program, "square", &err);
|
|
CheckCLError(err);
|
|
|
|
clInputBuffer = cl::Buffer(*context, CL_MEM_READ_ONLY, sizeof(float) * data_size, NULL, &err);
|
|
queue->enqueueWriteBuffer(clInputBuffer,
|
|
true, // Blocking!
|
|
0, sizeof(float) * data_size, gpuResult.data());
|
|
|
|
// Allocate the output data
|
|
clResultBuffer = cl::Buffer(*context, CL_MEM_WRITE_ONLY, sizeof(float) * data_size, NULL, &err);
|
|
|
|
// Set the kernel parameters
|
|
kernel.setArg(0, clInputBuffer); // kernel FV paraméterei sorrendben
|
|
kernel.setArg(1, clResultBuffer);
|
|
|
|
// Enqueue the kernel
|
|
queue->enqueueNDRangeKernel(kernel,
|
|
cl::NullRange, // Indexek nem eloffszetelve
|
|
cl::NDRange(data_size, 1), // Minden elemet egy szál
|
|
cl::NullRange, // Workgroup méret? - ez az auto, ha nem indul, 1024-re, onnan csökkent, amig elindul
|
|
NULL, //
|
|
Event); // Õ jlezi hogy vége, lsd lent
|
|
}
|
|
|
|
void Square::cpu_compute()
|
|
{
|
|
for (size_t index = 0; index < data_size; index++) {
|
|
cpuResult.push_back(sourceData[index] * sourceData[index]);
|
|
}
|
|
}
|
|
|
|
void Square::collect_results(cl::CommandQueue* queue) {
|
|
queue->enqueueReadBuffer(clResultBuffer, true, 0, sizeof(float) * data_size, gpuResult.data());
|
|
}
|
|
|
|
bool Square::validate_results() {
|
|
for (size_t index = 0; index < data_size; index++) {
|
|
if (cpuResult[index] != gpuResult[index]) {
|
|
std::cout << "Wrong result at [" << index << "]: " << gpuResult[index] << "!=" << cpuResult[index] << std::endl;
|
|
return false;
|
|
}
|
|
}
|
|
return true;
|
|
}
|
|
|
|
std::string Square::description()
|
|
{
|
|
return std::string("Square (data_size = " + std::to_string(data_size) + ")" );
|
|
}
|
|
|