Rockthesmurf
Advanced Member
The PS2 is indeed a beast to emulate, the architecture provides many challenges, some of which involve very timing dependent techniques (using specific number of NOPs in a DMA stream in order to wait for the DMA Controller to round robbin to start something else etc.). There are many challenges to be overcome, and then there is the fact the PS2 had a decent fill rate, something which mobile GPUs still typically do not. Some numbers:
GTX 660M : 13,400,000,000 pixels per second < Decent modern laptop GPU
PS2 : 2,400,000,000 pixels per second < The GS on the PS2
Pandora : 200,000,000 pixels per second < The Pandora PowerVR SGX530
This is purely pixel fill rate for the most basic pixels (untextured, etc.) and isn't really representative of what typically get drawn frame by frame, but it at least gives some idea where we are hardware wise. I wouldn't like to stick my neck out and guess what specification would be needed for PS2 emulation on the next Pandora, but even with the hardware, the software would be a problem. The work Exophase has done for DS is amazing (as is the dynarec work from Ari) and I don't want to be seen to down playing their amazing work (well beyond the capabilities of most), however my understanding is that PS2 is on another level; the idea of anyone in the PS2 community sitting down and writing a new PS2 emulator that kicks ass (like Exophase did with DS) that beats existing ones doesn't really seem realistic. Of course this is a double win for me; either I'm right, or if I'm proven wrong, I'd also be just as happy
GTX 660M : 13,400,000,000 pixels per second < Decent modern laptop GPU
PS2 : 2,400,000,000 pixels per second < The GS on the PS2
Pandora : 200,000,000 pixels per second < The Pandora PowerVR SGX530
This is purely pixel fill rate for the most basic pixels (untextured, etc.) and isn't really representative of what typically get drawn frame by frame, but it at least gives some idea where we are hardware wise. I wouldn't like to stick my neck out and guess what specification would be needed for PS2 emulation on the next Pandora, but even with the hardware, the software would be a problem. The work Exophase has done for DS is amazing (as is the dynarec work from Ari) and I don't want to be seen to down playing their amazing work (well beyond the capabilities of most), however my understanding is that PS2 is on another level; the idea of anyone in the PS2 community sitting down and writing a new PS2 emulator that kicks ass (like Exophase did with DS) that beats existing ones doesn't really seem realistic. Of course this is a double win for me; either I'm right, or if I'm proven wrong, I'd also be just as happy
Last edited by a moderator:
