Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Batch/Many FFT #140

Open
lashgar opened this issue Oct 16, 2023 · 3 comments
Open

Batch/Many FFT #140

lashgar opened this issue Oct 16, 2023 · 3 comments

Comments

@lashgar
Copy link

lashgar commented Oct 16, 2023

Does VkFFT support performing multiple FFTs in one call? A functionality similar to cufftPlanMany: https://docs.nvidia.com/cuda/cufft/index.html?highlight=cufftPlanMany#function-cufftplanmany

@DTolm
Copy link
Owner

DTolm commented Oct 16, 2023

Hello,

Yes, it is supported. There are also multiple ways to achieve these.

  1. You can use the numberBatches parameter, which is a straightforward number of consecutive systems in memory to do N-dimensional FFTs on. (See section 6.3 in documentation)
  2. You can select the higher number of dimensions and use omitDimension[i] parameter to disable the dimensions from performing the FFT on. With it, you can do the inner-most batching and have multiple types of batches with different strides. (See section 6.5 in documentation)

istride, idist and other parameters can be set with custom strides - see section 6.4

Best regards,
Dmitrii

@yves-surrel
Copy link

Is there a plan to allow omitDimension for r2c transforms?

@DTolm
Copy link
Owner

DTolm commented Sep 4, 2024

It is planned, but not in the nearest releases. The focus right now is the callbacks and inlining.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants