Hello my name is Daniel Hanrahan. Do you think my games should have optional calls to assembly functions for certain CPUs and GPUs in order to reach maximum possible performance and use parts/functions of the CPUs and GPUs that are not standard, but it’s use for them would for tasks that is not perfect for the standard parts of CPUs and GPUs if possible, for example: instead of using the standard parts/functions of the z80 for the randomization you use the refresh register instead. Let me be clear my games have good performance.
Link to my games: https://daniel-hanrahan-tools-and-games.github.io/
This is called compute dispatching and is super common. I have done a whole bunch of DSP implementations where you use a CUDA or AVX kernel depending on availability. Or you dispatch to standard library or even Python kernel otherwise.
Do you think compute dispatching is worth it for my video games.