cufft一維複雜到複雜的變換

Question

我想寫一個簡單的代碼fft 1d變換使用cufft庫。

我以cufft庫教程（link）上的代碼爲例，但在變換之前和逆變換之後的數據不是相同的。爲什麼？

這是輸出：

[1DCUFFT] starting... 

[1DCUFFT] is starting... 
Transforming signal cufftExecC2C 
Transforming signal back cufftExecC2C 
first : 0.840188 0.000000 after 2.020520 -2.465333 
first : 0.394383 0.000000 after 2.690347 -2.105700 
first : 0.783099 0.000000 after 3.155561 -1.491952 
first : 0.798440 0.000000 after 3.309971 -0.761629 
first : 0.911647 0.000000 after 3.139909 -0.092953 
first : 0.197551 0.000000 after 2.734147 0.355526 
first : 0.335223 0.000000 after 2.256154 0.501509 
first : 0.768230 0.000000 after 1.887422 0.369626 
first : 0.277775 0.000000 after 1.762471 0.083391 
first : 0.553970 0.000000 after 1.920150 -0.179526 
first : 0.477397 0.000000 after 2.289955 -0.252255 
first : 0.628871 0.000000 after 2.718623 -0.039409 
first : 0.364784 0.000000 after 3.026649 0.449287 
first : 0.513401 0.000000 after 3.072947 1.104457 
first : 0.952230 0.000000 after 2.803322 1.758350 
first : 0.916195 0.000000 after 2.265463 2.245056 

這是我的源代碼，教程示例建議劃分爲數據集大小輸出到具有相同的值。

// includes, system 
#include <stdlib.h> 
#include <stdio.h> 
#include <string.h> 
#include <math.h> 

// includes, project 
#include <cufft.h> 
#include <cutil_inline.h> 
#include <shrQATest.h> 
void runTest(int argc, char** argv); 

// The filter size is assumed to be a number smaller than the signal size 
#define SIGNAL_SIZE  16 

//////////////////////////////////////////////////////////////////////////////// 
// Program main 
//////////////////////////////////////////////////////////////////////////////// 
int main(int argc, char** argv) 
{ 
    runTest(argc, argv); 
} 

//////////////////////////////////////////////////////////////////////////////// 
//! Run a simple test for CUDA 
//////////////////////////////////////////////////////////////////////////////// 
void runTest(int argc, char** argv) 
{ 

    printf("[1DCUFFT] is starting...\n"); 

    if(cutCheckCmdLineFlag(argc, (const char**)argv, "device")) 
     cutilDeviceInit(argc, argv); 
    else 
     cudaSetDevice(cutGetMaxGflopsDeviceId()); 
    cufftComplex* h_signal=(cufftComplex*)malloc(sizeof(cufftComplex) * SIGNAL_SIZE); 
    // Allocate host memory for the signal 
    //Complex* h_signal = (Complex*)malloc(sizeof(Complex) * SIGNAL_SIZE); 
    // Initalize the memory for the signal 
    for (unsigned int i = 0; i < SIGNAL_SIZE; ++i) { 
     h_signal[i].x = rand()/(float)RAND_MAX; 
     h_signal[i].y = 0; 
    } 




    int mem_size = sizeof(cufftComplex) * SIGNAL_SIZE; 

    // Allocate device memory for signal 
    cufftComplex* d_signal; 
    cutilSafeCall(cudaMalloc((void**)&d_signal, mem_size)); 

    // Copy host memory to device 
    cutilSafeCall(cudaMemcpy(d_signal, h_signal, mem_size, 
           cudaMemcpyHostToDevice)); 



    // CUFFT plan 
    cufftHandle plan; 
    cufftSafeCall(cufftPlan1d(&plan, mem_size, CUFFT_C2C, 1)); 

    // Transform signal 
    printf("Transforming signal cufftExecC2C\n"); 
    cufftSafeCall(cufftExecC2C(plan, (cufftComplex *)d_signal, (cufftComplex *)d_signal, CUFFT_FORWARD)); 


    // Transform signal back 
    printf("Transforming signal back cufftExecC2C\n"); 
    cufftSafeCall(cufftExecC2C(plan, (cufftComplex *)d_signal, (cufftComplex *)d_signal, CUFFT_INVERSE)); 

    // Copy device memory to host 
    cufftComplex* h_inverse_signal = (cufftComplex*)malloc(sizeof(cufftComplex) * SIGNAL_SIZE);; 
    cutilSafeCall(cudaMemcpy(h_inverse_signal, d_signal, mem_size, 
           cudaMemcpyDeviceToHost)); 

    for(int i=0;i< SIGNAL_SIZE;i++){ 
     h_inverse_signal[i].x= h_inverse_signal[i].x/(float)SIGNAL_SIZE; 
     h_inverse_signal[i].y= h_inverse_signal[i].y/(float)SIGNAL_SIZE; 

     printf("first : %f %f after %f %f \n",h_signal[i].x,h_signal[i].y,h_inverse_signal[i].x,h_inverse_signal[i].y); 
    } 



    //Destroy CUFFT context 
    cufftSafeCall(cufftDestroy(plan)); 

    // cleanup memory 
    free(h_signal); 

    free(h_inverse_signal); 
    cutilSafeCall(cudaFree(d_signal)); 
    cutilDeviceReset(); 
} 

Answer 1

你需要改變：

cufftSafeCall(cufftPlan1d(&plan, mem_size, CUFFT_C2C, 1)); 

到：

cufftSafeCall(cufftPlan1d(&plan, SIGNAL_SIZE, CUFFT_C2C, 1)); 

（cufftPlan1d需要的數據點，而不是內存大小中的字節數！）