Re: Non-local means denoising
On Sat, 25 Jan 2014 23:49:29 +1100, Juan Nunez-Iglesias wrote:
I'm personally very interested, yes! Regarding scikit-image, I'm not sure whether Stéfan would be happy to go down the GPU route (and CUDA more specifically), but my feeling is that a very large chunk of scientific computing is going in that direction, so it would be good to have at least an optional dependency there...
We've considered this before (recall discussions about multiple backends), but in the end found that it is not viable to maintain multiple implementations of the same code (at least not for our small team). I am putting my bets on tools like Numba, Parakeet, etc. to save us from Python's slowness. Stéfan
On Sun, Jan 26, 2014 at 5:44 AM, Stéfan van der Walt <stefan@sun.ac.za>wrote:
We've considered this before (recall discussions about multiple backends), but in the end found that it is not viable to maintain multiple implementations of the same code (at least not for our small team).
I do remember these, but it's a worthwhile conversation to bring up every so often as tools mature and more people contribute. To be honest I know next to nothing about CUDA and GPU programming, I just see the success stories fly by. =) I am putting my bets on tools like Numba, Parakeet, etc. to save us from
Python's slowness.
Yeah these are pretty exciting. =) Juan.
participants (2)
-
Juan Nunez-Iglesias -
Stéfan van der Walt