0
私はこのコードを実行して推力のあるstable_sortとユーザー定義の演算子を使ってIPを比較します。 このコードは50000未満のIPの配列で動作しますが、大きな配列に対してはメモリエラーが発生します。 メモリの場所のエラー：大きな配列とユーザー定義の比較演算子を使用すると推力:: stable_sort
がどのように大きなアレイとこの問題を解決するためのメモリ位置での推力::システム:: SYSTEM_ERROR：私が得た#include "cuda_runtime.h"
#include "device_launch_parameters.h"
#include <thrust/host_vector.h>
#include <thrust/device_vector.h>
#include <thrust/sort.h>
#include <stdio.h>
#include <time.h>
#include <device_functions.h>
template<typename T>
struct vector_less
{
typedef T first_argument_type;
typedef T second_argument_type;
typedef bool result_type;
__host__ __device__ bool operator()(const T &lhs, const T &rhs) const {
if (lhs[0] == rhs[0])
if (lhs[1] == rhs[1])
if (lhs[2] == rhs[2])
return lhs[3] < rhs[3];
else
return lhs[2] < rhs[2];
else
return lhs[1] < rhs[1];
else
return lhs[0] < rhs[0];
}
};
__global__ void prepare_ips_list(unsigned char ** dev_sorted_Ips, unsigned char * ip_b1, unsigned char * ip_b2, unsigned char * ip_b3, unsigned char * ip_b4, unsigned int searchedIpsSize)
{
int thread = threadIdx.x + blockIdx.x * blockDim.x;
if (thread < searchedIpsSize)
{
dev_sorted_Ips[thread] = new unsigned char[4];
dev_sorted_Ips[thread][0] = ip_b1[thread];
dev_sorted_Ips[thread][1] = ip_b2[thread];
dev_sorted_Ips[thread][2] = ip_b3[thread];
dev_sorted_Ips[thread][3] = ip_b4[thread];
}
}
int main()
{
const int size = 1000000;
unsigned char * ip_b1 = new unsigned char[size];
unsigned char * ip_b2 = new unsigned char[size];;
unsigned char * ip_b3 = new unsigned char[size];;
unsigned char * ip_b4 = new unsigned char[size];;
unsigned char * dev_ip_b1;
unsigned char * dev_ip_b2;
unsigned char * dev_ip_b3;
unsigned char * dev_ip_b4;
unsigned char ** dev_sortedIps;
for (int i = 0; i < size; i++)
{
ip_b1[i] = rand() % 240;
ip_b2[i] = rand() % 240;
ip_b3[i] = rand() % 240;
ip_b4[i] = rand() % 240;
}
cudaError_t cudaStatus;
cudaStatus = cudaSetDevice(0);
if (cudaStatus != cudaSuccess) {
fprintf(stderr, "cudaSetDevice failed! Do you have a CUDA-capable GPU installed?");
goto Error;
}
cudaStatus = cudaMalloc((void**)&dev_ip_b1, size * sizeof(unsigned char));
if (cudaStatus != cudaSuccess) {
fprintf(stderr, "cudaMalloc failed!");
goto Error;
}
cudaStatus = cudaMemcpy(dev_ip_b1, ip_b1, size * sizeof(unsigned char), cudaMemcpyHostToDevice);
if (cudaStatus != cudaSuccess) {
fprintf(stderr, "cudaMemcpy failed!");
goto Error;
}
cudaStatus = cudaMalloc((void**)&dev_ip_b2, size * sizeof(unsigned char));
if (cudaStatus != cudaSuccess) {
fprintf(stderr, "cudaMalloc failed!");
goto Error;
}
cudaStatus = cudaMemcpy(dev_ip_b2, ip_b2, size * sizeof(unsigned char), cudaMemcpyHostToDevice);
if (cudaStatus != cudaSuccess) {
fprintf(stderr, "cudaMemcpy failed!");
goto Error;
}
cudaStatus = cudaMalloc((void**)&dev_ip_b3, size * sizeof(unsigned char));
if (cudaStatus != cudaSuccess) {
fprintf(stderr, "cudaMalloc failed!");
goto Error;
}
cudaStatus = cudaMemcpy(dev_ip_b3, ip_b3, size * sizeof(unsigned char), cudaMemcpyHostToDevice);
if (cudaStatus != cudaSuccess) {
fprintf(stderr, "cudaMemcpy failed!");
goto Error;
}
cudaStatus = cudaMalloc((void**)&dev_ip_b4, size * sizeof(unsigned char));
if (cudaStatus != cudaSuccess) {
fprintf(stderr, "cudaMalloc failed!");
goto Error;
}
cudaStatus = cudaMemcpy(dev_ip_b4, ip_b4, size * sizeof(unsigned char), cudaMemcpyHostToDevice);
if (cudaStatus != cudaSuccess) {
fprintf(stderr, "cudaMemcpy failed!");
goto Error;
}
cudaStatus = cudaMalloc((void**)&dev_sortedIps, size * sizeof(unsigned char *));
if (cudaStatus != cudaSuccess) {
fprintf(stderr, "cudaMalloc failed!");
goto Error;
}
int resetThreads = size;
int resetBlocks = 1;
if (size > 1024)
{
resetThreads = 1024;
resetBlocks = size/1024;
if (size % 1024 > 0)
resetBlocks++;
}
prepare_ips_list << <resetBlocks, resetThreads >> >(dev_sortedIps, dev_ip_b1, dev_ip_b2, dev_ip_b3, dev_ip_b4, size);
thrust::device_ptr<unsigned char *> sorted_list_ptr1(dev_sortedIps);
thrust::stable_sort(sorted_list_ptr1, sorted_list_ptr1 + size, vector_less<unsigned char *>());
cudaStatus = cudaGetLastError();
if (cudaStatus != cudaSuccess) {
fprintf(stderr, "launch failed: %s\n", cudaGetErrorString(cudaStatus));
goto Error;
}
// cudaDeviceSynchronize waits for the kernel to finish, and returns
// any errors encountered during the launch.
cudaStatus = cudaDeviceSynchronize();
if (cudaStatus != cudaSuccess) {
fprintf(stderr, "cudaDeviceSynchronize returned error code %d after launching !\n", cudaStatus);
goto Error;
}
return 0;
Error:
cudaFree(dev_ip_b1);
cudaFree(dev_ip_b2);
cudaFree(dev_ip_b3);
cudaFree(dev_ip_b4);
cudaFree(dev_sortedIps);
}
エラーは次のとおりです： のMicrosoft C++例外 ここで私が使用したコードはありますか？ 私はこのソートを達成するために、別の手法を使用して、マージする部品の分割やソートなどを行うべきですか？
[タグ：c]ではありません。ここ –