Ah I missed this.
It really depends on thread count for mine and how complex the polygons are shaded. Such as lightmaps, textures, shadows, etc and especially size! If each polygon is rasterized at 1 pixel in size with flat color, then that kind of number on that kind of hardware (threaded) isn't unreasonable
Practically, I am seeing faster speeds and higher resolutions than Quake 1 (with appropriate occlusion culling) on a fairly moderate i3 in my laptop. This is likely down to multiple threads (I didn't do any assembly, etc). But my code is open-source so you are very free to benchmark it yourself.
You might be interested in benchmarking against Bellard's TinyGL (https://bellard.org/TinyGL/). This one is fairly well optimized. Mine can generally produce identical output with a higher speed but again, only due to the multithreading.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.