I've actually been looking over the source code, and if you wouldn't mind I would like to recommend a few improvements that you might make to PCSX. There are a few more optimizations that can be done to the recompiler to increase its speed even more, here's a list of possible ideas:
- Caching of pre-compiled x86 code so that when loops are being processed the code doesn't have to be repeatedly regenerated.
- You could also consider using circular buffers for loop-processing so that when a jump instruction is executed to previous code that has already been recompiled it will be immediately available in this recompiled form, this may take some r3000a binary analysis though.
- Just a note, keeping track of these would require a 2D array that is written/read from based on the memory location of the translated instruction. This way you can check to see if you have already compiled the code at that location thus avoiding recompiling it again if it is not necessary.
These are just some thoughts from looking through the source code. Best of luck to you, and hope you make PCSX better than it already is.
-zenogais