Almar Klein
Almar Klein
Status: all the chunking stuff is implemented and tested. Next up is doing some benchmarking so I can tweak the implementation and chunk parameters for performance.
The `dim` arg is meant to remove this ambiguity. It can indeed fix it for channels, but I realize now that it cannot really disambiguate for `N`, nor any collapsed...
> I think that when users explicitly specify the size of a Texture, we should disregard other conditions and adhere to the user-specified size, adjusting the data to fit this...
Created benchmarks for texture. For now, the chunking eum ... negatively affects performance. I'll try to find out whether this is just due to chunking too much, or due to...
Pfew ... 😌
Run this: https://github.com/pygfx/pygfx-benchmarks/blob/main/benchmarks/bm_texture.py
Yeah, manually keeping this list up-to-date is a pain and likely to be forgotten. We have similar lists in the `__init__.py` of the subpackages. A script might be a good...
I'm cleaning up my inbox, and realize I had missed this earlier. Anyway, yeah, I think I'd add logic to the pointer-move handler, and if the pointer has moved more...
I can see your point and how this might be useful. Are all the classes in that list abstract (not implementing anything)? The only caveat I see is that you...
Nice! > Pretty hard to benchmark though... Let's try though! E.g. https://github.com/pygfx/pygfx-benchmarks/blob/main/benchmarks/bm_points.py