will the team keep making the emulator more efficient

[Blyss Sarania]

To fix a things like mipmapping GSdx would need an entire rewrite. At this point getting more accuracy is goin to require a major overhaul, as for everyone saying that there should be a competing emulator, there aren't a whole lot of people that understand the ps2 like the devs. I really don't think the accuracy will increase much more.

I think the accuracy is less of a problem at this point than the efficiency.

And as for a lot of people not understanding it like the PCSX2 devs - that's the point I was making. Potential new devs can learn almost everything they need to know from the PCSX2 source. For the PCSX2 devs, if they had a problem, they had to figure out the hardware and write the code for it all on their own. A new emulator can just look at PCSX2 code and say "Oh, that's how it's done. It think it would be faster this way."

[nini]

You guys do understand the difference in complexity between a PS2 and a PSP right?? Just saying, I also agree the advancements on PPSSPP are quite astronomical as of lately, but dont 2 AFAIK is a more complex systems with more synchronization needed between its multiple Processing Units... I don't think the developers slack off, I think what is left to fix is quite complex...

The fact remains that the ppsspp project is improving performance dramatically with every update. In the last post on the official website they said that the next update would be dedicated entirely to performance and optimization. Why dont the pcsx2 team do the same?

[NarooN]

Yeah, there won't be any magical advancements made with PCSX2 any time soon. We might get a couple new features at some point (what, exactly, I dunno) but PCSX2 as it stands is pretty much sort of at a wall. There's only a handful of active devs and they all have actual lives and families they have to tend to as well. The only way something huge would happen is if a huge wave of talented devs joined the project out of nowhere, but that most likely won't happen any time soon.

The thing about PCSX2 is that it has been a WIP for over a decade now. That code has been modified and optimized as well as it could've been since then. For PS2 games to be emulated at fullspeed on such lower end hardware would require an entirely new emu. That would be awesome, though there doesn't seem to be much interest in PS2 emu nowadays. I guess most people who were interested in the past looked at how complex the PS2's uArch was, saw how far PCSX2 had gotten, and decided "welp, f*** that, kthxbai."

[Nobbs66]

Yeah, there won't be any magical advancements made with PCSX2 any time soon. We might get a couple new features at some point (what, exactly, I dunno) but PCSX2 as it stands is pretty much sort of at a wall. There's only a handful of active devs and they all have actual lives and families they have to tend to as well. The only way something huge would happen is if a huge wave of talented devs joined the project out of nowhere, but that most likely won't happen any time soon.

The thing about PCSX2 is that it has been a WIP for over a decade now. That code has been modified and optimized as well as it could've been since then. For PS2 games to be emulated at fullspeed on such lower end hardware would require an entirely new emu. That would be awesome, though there doesn't seem to be much interest in PS2 emu nowadays. I guess most people who were interested in the past looked at how complex the PS2's uArch was, saw how far PCSX2 had gotten, and decided "welp, f*** that, kthxbai."

I'm not convinced a new emulator would bring any speed increases. The devs have optimised the code to death and I don't know how much more you could even possibly do.

[NarooN]

There could be increases with completely fresh code, even if it's loosely-based on the PCSX2 source. The devs have already told people before (sarcastically) that they could always rewrite PCSX2 from scratch or add to the source themselves when people asked for fairly ridiculous things ("plz ad spprt 4 16-coar procczzrs plz", etc.) For example, PCSX2 has pretty much always been Intel-biased since the get-go. A totally new PS2 emu, provided it was a decently-sized team with very talented members with lots of free time (pipe dream), they could code the emu in such a way that it leverages all modern processor uArchs to squeeze out the max performance from all chips.

This is all hypothetical stuff obviously. For now, the only way to get more performance (and the only way it will probably remain for the foreseeable future) is to simply upgrade your hardware.

[Blyss Sarania]

I'm not convinced a new emulator would bring any speed increases. The devs have optimised the code to death and I don't know how much more you could even possibly do.

Well, PCSX2 is mostly an LLE. HLE is usually faster, and so if that approach was taken, it definitely would be faster. Of course, HLE is less accurate.

BTW
LLE = low level emulation
HLE = High level emulation

[Nobbs66]

Well, PCSX2 is mostly an LLE. HLE is usually faster, and so if that approach was taken, it definitely would be faster. Of course, HLE is less accurate.

BTW
LLE = low level emulation
HLE = High level emulation

If pcsx2 were a HLE it would probably have more compatibility issues than it's worth. Just because you would normaly gain speed it may not hold true for the ps2 because of the complexity of the architecture. NFS happens to be an example of this. If the accuracy was improved NFS might run faster than it does currently.

[Blyss Sarania]

If pcsx2 were a HLE it would probably have more compatibility issues than it's worth. Just because you would normaly gain speed it may not hold true for the ps2 because of the complexity of the architecture. NFS happens to be an example of this. If the accuracy was improved NFS might run faster than it does currently.

Well, this is all useless babbling on our part anyway lol. PCSX2 is what it is. It's not gonna magically change into something else. The only thing that really bothers me is a couple of GSDX issues, but I'm gonna try to correct those myself.

[Nobbs66]

Well, this is all useless babbling on our part anyway lol. PCSX2 is what it is. It's not gonna magically change into something else. The only thing that really bothers me is a couple of GSDX issues, but I'm gonna try to correct those myself.

If your working fixing issues can you implement mipmapping in hardware mode.

[Blyss Sarania]

If your working fixing issues can you implement mipmapping in hardware mode.

Haha, I'm sure that's beyond me for now. Really what I want to do is change the texture filtering to something better than bilinear. Just to write something that will work shouldn't be hard, of course making it fast is another story.

Mipmapping works in software though, I wonder why it can't be ported to hardware?