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

Refacto CUDA files #609

Open
simon-riou opened this issue Oct 14, 2024 · 4 comments
Open

Refacto CUDA files #609

simon-riou opened this issue Oct 14, 2024 · 4 comments
Assignees
Labels
cleanup Makes the code cleaner non-urgent non-urgent issues

Comments

@simon-riou
Copy link
Collaborator

Structure might not always be the best
Useless includes
Dead code
Unused function
Bad comments
...

@simon-riou simon-riou added non-urgent non-urgent issues cleanup Makes the code cleaner labels Oct 14, 2024
@simon-riou simon-riou self-assigned this Oct 14, 2024
@simon-riou
Copy link
Collaborator Author

noTban pushed a commit that referenced this issue Oct 16, 2024
@simon-riou
Copy link
Collaborator Author

simon-riou commented Oct 16, 2024

Todo list of CUDA files

  • apply_mask.cu
  • composite.cu
  • convolution.cu
  • angular_spectrum.cu
  • frame_reshape.cu
  • input_filter.cu
  • percentile.cu
  • rgb.cu
  • stft.c
  • texture_update.cu
  • tools_compute.cu
  • tools_hsv.cu
  • transforms.cu
  • chart.cu
  • contrast_correction.cu
  • fresnel_transform.cu
  • filter2D.cu
  • hsv.cu
  • map.cu
  • reduce.cu
  • shift_corners.cu
  • test_reduce.cu (What is the purpose of this file ?)
  • tools.cu
  • tools_conversion.cu
  • tools_unwrap.cu
  • unpack.cu
  • cuda_memory

@Hpn4
Copy link
Collaborator

Hpn4 commented Oct 17, 2024

You can delete:

  • frame_reshape.cc and frame_reshape.hh
  • in reduce functions: reduce_min, reduce_max, gpu_reduce and delete duplicate functions that are in reduce.cc and in reduce.cuhxx (you can basically delete reduce.cc)
  • in map functions: map_divide
  • in shift_corners functions that have 6 paramètres (only the 5 parameters one are called)

@simon-riou
Copy link
Collaborator Author

simon-riou commented Oct 17, 2024

https://developer.nvidia.com/blog/using-cuda-warp-level-primitives/

I use it in kernel_apply_mapped_zone_sum in chart.cu to avoid too many calls of atomic op

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
cleanup Makes the code cleaner non-urgent non-urgent issues
Projects
None yet
Development

No branches or pull requests

2 participants