M1 Pro/Max uses unified memory for both CPU and GPU, does it simplify the structure of GPU-accelerated parallel computing? Or is it still using the same code like AMD GPU?
Is there any change in GPU Computing (MSL) when dealing with M1 Pro/Max?
It depends on what context. From a Metal API perspective, you shouldn't need to change much (if any) of your API calls to support Apple Silicon GPUs. You also shouldn't need to change your shader code, unless you are planning to take advantage of GPU-specific features. The Metal compiler will convert your shader into different machine code depending on if you are targeting Apple Silicon, Intel, or AMD GPUs, but this is more of an implementation detail.
Unified memory does simply everything a bit, since it reduces synchronization between CPU and GPU memory that exists on other GPUs. You can learn more about Metal on Apple Silicon Macs in these videos:
https://developer.apple.com/videos/play/tech-talks/10580/
https://developer.apple.com/videos/play/tech-talks/110147/
https://developer.apple.com/videos/play/tech-talks/10859/