in software rendering which of these scenario’s would be faster?
320 polys, or
64 untextureded polys, and
6 32 bit rgba textures shared between 32 polys,
You’ll have to be more specific than that, but not not just code what you’re thinking of and find out? However my guess is that the textured polys (properly optimized) would be faster, both doom II and quake made extensive use of textures but were very low-poly compared with games designed for hardware acceleration. I suppose it you’re spending all your clock cycles rasterizing and blitting to the framebuffer it doesn’t matter so much where the pixels come from as how many you draw.