Since pretty much everyone who plays games is gonna have a graphics card, is it even worth exploring CPU graphics engines?

Turbo14 said:
is it even worth exploring CPU graphics engines?

It is worth for me for research purposes. I have tried a lot of different algorithms for GI and found it easier and more comfortable to do it on CPU.
Then, after making a choice, i have ported it to GPU compute.

Also, i have a raytracer which i can use to experiment with surface ‘shaders’ without a need to work on GPU which always is cumbersome.
Again, port it to GPU when done. (But probably this approach can't beat productivity you would get from using U-engine with its node based shader graphs ofc.)

What was the last game with 3D graphics on CPU? Probably Outcast 20 years ago.

I think it is important, as it ties into the fundamentals, and fundamentals tie into really low-level things down the line. Even doing a simple exercise like Bresenham's rasterization can teach you more than just “knowing” how rasterization works on a GPU. Or doing manual clipping.