Curvelet Transform Acceleration – GPU or Multithreading Support

[LessandroValente]

Hello,

I am currently conducting some seismic experiments and testing the Curvelet transform (Curvelops). However, I have not been able to find any support for running it on a GPU, and I suspect that such acceleration may not be available.

As an alternative, I attempted to parallelize the process using multiple threads, but encountered a PyCapsule error during execution.

Could anyone provide guidance on whether it is possible to accelerate the Curvelet operator using any optimization techniques, such as GPU support or multithreading?

Thank you in advance for your help.

[cako]

Hi @LessandroValente, I sent you an email with more details but long story short, it's not possible with Curvelops. If you want to check out https://github.com/cako/curvelets for that, let me know and we can chat.