NVIDIA / NVIDIA/CUDALibrarySamples
cufft_lto_ea example does not work under windows
@mferreravila is already working on this.
Since Jun 27, 2025.
- Dominant language
- Cuda
- Stars
- 2.5k
- Forks
- 478
- PR merge metrics
- No merged PRs in 30d
Description
hello,
I am trying to use the LTO EA version of cufft, but under windows 11 the r2c_c2r_lto_callback example output:
Example failed in cuFFT API on line 99 with error 8
^^^^ the above error message ('Binary format for key...') can be safely ignored, it will be fixed in a future release
Transforming signal cufftExecR2C
Transforming signal cufftExecC2R
Example failed in cuFFT API on line 108 with error 10
Transforming reference cufftExecR2C
Transforming reference cufftExecC2R
L2 error: 1.003510e+00
It seems that cufftplan creation:
CHECK_ERROR(cufftMakePlan1d(inverse_plan_cb, signal_size, CUFFT_C2R, batches, &work_size));
issues an error. And when I comment out the the line contains cufftXtSetJITCallback, the error is gone.
What is more, to avoid a stack overflow error, I make the following changes to the example:
_//Padded array for in-place transforms
float input_signals[batches][2 * complex_signal_size] = {};
float output_signals[batches][2 * complex_signal_size];
float reference[batches][2 * complex_signal_size];
->
_static float input_signals[batches][2 * complex_signal_size] = {};
static float output_signals[batches][2 * complex_signal_size];
static float reference[batches][2 * complex_signal_size];__
And my cuda version is 12.4.1 with cufft replace by the ea version cufft 11.1.0., and test with rtx2070 maxq.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Assessment
This issue has not been assessed yet.