Home Artists Posts Import Register

Content

The caching system didn't work as expected. It worked as far as optimizing for full computation of texel coordinates per-pixel, but didn't work for computing steps between texel coordinates.

Anyway, while working on it I've figured out that the s & t "adjust" variables can be precomputed into the dithering of the coordinates to eliminate one sum per coordinate every time the perspective is corrected. This gave a nice little performance increase to the rasterizer, but only works on texel coordinates; to make it work with the coordinates of the lighting, the luxel coordinates must be computed as power-of-2, which is a task that's on my to-do list.

Now I'm trying to implement the same approach on liquid surfaces, which is trickier due to the way the waves are computed. But it should work.

I'm also busy trying to get a lawyer in order to open a lawsuit to sell my stake in my late mother's house, which would fix my financial ruin. I've talked to a few lawyers already, and am considering the options.

Comments

No comments found for this post.