Why shouldn't it be?Originally Posted by JKKDARK
Why shouldn't it be?Originally Posted by JKKDARK
Exophase:
Be sure to check PMs. ;P
I forgot to mention register mapping. How many registers are available for you to map to on a MIPS32 host when doing gpSP? Right now I've mapped registers $10 - $23. I'm assuming I can use more but haven't gone through tweaking it yet.
For caching I've been trying to work on a segmented translation cache system, but it, as you said it would, is failing. Having over 16MB for a translation cache on the V43 makes this a little less of a worry, but on MIPS gadgets that dont have as much RAM, it's sad.
One thing I have done though with the R3000A -> ARM dynarec is make it "threaded" in the sense that the recompiled function is able to jump into itself and execute the compile function on a branch (the tailend). That way it never has to leave the recompiled function, and some interesting optimizations can be possible.
One thing I'm looking into ( thx Tinnus :P ) is reducing the need to recompile LUI, ORI combinations for MIPS way of loading a 32bit immediate. Since I map immediates to registers, and store the contents in a struct, I can technically do it during the compile function. Might be tricky though.
How are you going about clearing the icache for your dynarec? For MIPS there's the cache instruction, but it requires kernel mode. For ARM there's MCR but it too requires kernel mode. I heard there's a SWI for doing a icache clear though? Would be better to be able to do a hit invalidate instead of index invalidate of course. Feel free to PM as needed. We might be able to help each other out. :P
it seems that you don't know what happened with this emulator recently..Originally Posted by Festizzio
man so much conplaining. Yoshi made a mistake big woop. Everyone makes a mistake yoshi can fix it by posting a source code. Yoshi you are still a great coder. all people have to agree that it takes talent make an emulator of this magnitude to work. While some people hates his work on his emulator but i am sure he will make the legal and will be known as one of the best coder in the psp scene
Yoshi is a ****** name one program he made himself and did not just steal code from someone else.Originally Posted by psman
A friend of mine agreed to translate the front page of pspgen.com for us. His French is a little rusty and I might be editing this at his will later but it should be a lot better than Babelfish. Here you go:
If this is all true I'd say the blame has been pretty seriously misplaced with all this. Sounds like pspgen didn't handle this as well as they could have, more than yoshihiro doing anything wrong...Originally Posted by pspgen.com English translation
It doesn't. zodttd ported PCSX to the Tapwave Zodiac in literally one night. Although that was just a first shot, and psx4all was done a lot of time later (with the knowledge and experience zodttd gathered from the port).Originally Posted by psman
You'd be surprised, but a PS1 emulator is not very complicated. Less complicated than a SNES one, and maybe even less complicated than a NES one (due to all the mapper stuff). And certainly less complicated than an Atari 2600 emulator
A system being more powerful doesn't mean it's harder to emulate or to "get working" in another platform. The 2600 is a very hard system to emulate due to it having tons of chips and tons of stuff done in hardware like collision detection.
so this Yoshi's emu is modifyed from some other person PSX project?
So did he give credit from the source he used, because I know there was a big fiasco about that.
There are currently 1 users browsing this thread. (0 members and 1 guests)
Bookmarks