Stack Overflow
Stack Overflow

Questions tagged [cufft]

cuFFT is a FFT library for CUDA enabled GPUs. Capabilities are similar to the FFTW library.

cuFFT is a FFT library for CUDA enabled GPUs. cuFFT provides functions to do various kinds of forward and reverse Fast Fourier Transforms including multidimensional transforms and batched transforms.

146 questions
-1
votes
1 answer

Unable to allocate CUDA device memory for cufftComplex data type

I am trying to allocate a cufftComplex array into memory on a CUDA device (GEFORCE GTX 1080) using the following code: cufftComplex *d_in, *d_out; int ds = sizeof(cufftComplex) * width * height; CUResult test_din = cuMemAlloc((void**)&d_in,…
DLH
  • 199
  • 11
-1
votes
1 answer

CuFFT Unknown Error

I have an array of 300,000 points and I want the fft of every 600 points. I'm attempting to use cufftPlanMany to execute, but I'm getting an unknown error here: cufftSafeCall(cufftPlanMany(&plan, rank, n, NULL, istride, idist, NULL, 1,1, CUFFT_C2C,…
Eddie E Massey III
  • 323
  • 2
  • 14
-2
votes
1 answer

Blockwise/Strided reduction using CUDA

TLDR: I am trying to write a GPU code that computes a blockwise reduction on an array. The input looks like [block_0, trash_0, block_1, trash_1, ..., block_n, trash_n], and I want to compute block_0 + block_1 + ... + block_n. Relevant sizes: each…
s769
  • 3
  • 3
-2
votes
1 answer

cudaMemcpy throws InvalidValue error when copying from device to host

I've been trying to implement a one dimensional FFT using cuFFT. An InvalidValue error is thrown and no meaningful results are produced. I've tried to ensure that each error is caught, and I believe that the cudaMemcpy from DeviceToHost causes the…
obscuresc
  • 1
  • 2
-2
votes
1 answer

How to measure fft and ifft time from the cuda::convolution function?

I am using the cuda::convolution::convolve to calculate the Gaussian convolution and I want to measure the time of the fft and ifft. But I don't know how to measure. I found the source code on the GitHub. I have no idea how to measure the time from…
Zaw
  • 15
  • 5
-2
votes
1 answer

CUDA: cufftExecR2C - Unnecessary memory copies

I am experimenting with cuda and observe that data is copied from host to device when I invoke cufftExecR2C(plan, src, dst); which I don't undertand since my src pointer is a valid handle to the device memory that I would like to transform. Before…
Martin1988a
  • 79
  • 7
-2
votes
1 answer

cuFFT looped FFT calls for bigger batch sizes

I'm currently trying to run my multiple FFT's in a loop to overcome the 128 million element max of the cuFFT plan. So, for example, I would run 128 million element runs in a loop. My program works just fine for a single FFT call, but looping doesn't…
EE_Guy
  • 113
  • 1
  • 1
  • 9
-2
votes
1 answer

1d complex to complex transform with cuda cufft library

i'm writing a simple code for complex to complex 1d FFT with cuda library "cufft". I've changed "cudaMalloc" with "malloc" because if i use cudaMalloc exe file not work and crashing! But this is not the primary problem. When i execute the program,…
Marco Jammajalla Mangraviti
  • 149
  • 3
  • 15
-2
votes
1 answer

Using cucomplex.h to manipulate cufft data

I'm new here. I'm working with FFT, and I need to make a simple code, but it's not working. I need to transform with cufft a sin(x) and turn back, but between the transforms, I need to multiply by 2 the result so that, when I turn back the result…
Túlio Alves
  • 13
  • 3
-2
votes
1 answer

cuFFT output not correct

I have a problem with this program: #include #include #include #include #include #include #define SIGNAL_SIZE 1024 int main(int argc, char **argv) { cudaEvent_t start,…
Marco
  • 23
  • 1
  • 4
-2
votes
1 answer

Evaluating the execution time of 1d cuFFT

The code snippet cudaEventRecord(start, 0); /* creates 1D FFT plan */ cufftPlan1d(&plan, NX, CUFFT_C2C, BATCH); /* executes FFT processes */ cufftExecC2C(plan, devPtr, devPtr, CUFFT_FORWARD); cudaEventRecord(stop, 0); …
Sreehari
  • 11
  • 2
1 2 3
…
9
10