I know youre talking raytracing here but (as i think there are people here who are doing much high efficiency stuff here) i wonder if there would be an option to discuss software cpu rasterization a bit?
I wrote simple soft scanline rasterizer in c and has a speed of about 30ms for 100thousand triangles model (with flat vertex shading) for 500p (500x500) image and about 70 ms for 1000p (1000x1000) for core2duo machine on 1 thread
Search for some hints hoe to speed up - also for some info values how more professional rasterizers are flying here
my code was very strightforward each trangle i just rotate, clip if even one vertex goes rear the view plane (so it is inprecise), scale by 1/z (so it is a bit wrong as far as i heard, thent test if thi is lying in screen rectangle in 2d, if so i rasterize it with scanline, thats all)
- could someone trained in optymizations give some advice? tnx
2 posts • Page 1 of 1
This is a pretty good paper (CUDA though, but still most of the concepts in there apply to CPU, too, incl. source code): https://research.nvidia.com/publication ... ation-gpus
Better you leave here with your head still full of kitty cats and puppy dogs.
Who is online
Users browsing this forum: No registered users and 0 guests