Psx4gp2x Public Beta 2 Released.

[Vollgasasi]

vudooman:
you know what the next step for you is right? order a gp2x immediatly!

 

[Shaun]

This is great news. Hopefully the great performance boost will mean the new version'll have sound? Also, please improve the menu, even if it means switching it back to the old style. ^_~ It's just that the current release's menu... I've found it impossible to navigate through it.

 

[X-Code]

I don't see much wrong with the menu, it's quite intuitive, but even so, I'm quite sure that it'll be improved over the next upcoming versions ... Maybe a GUI ? .

 

[Mortys]

Just succeed to make Psx4all work on my gp with Tobal 2. 9-10 fps with Interlace 1 and CPU 266. Can't wait to test next release. But I've got a question on 3D. How PSX4all draw the 3D ? Did they use an existing graphical engine (or something like that) or it's only software rendering ? And finally, is there a way to use hardware drawning?

 

[STTrife]

If you mean is there a hardware accelerator in the GP2X then: no
I've seen discussions about the possibilty to use the coprocessor or the GPU to accelerate floating point operations which could be considered hardware 3d acceleration. Some people believe it's possible but wheter it's possible or not, it's very difficult I think.
The real hardware guru's would have to take a look if it's possible. I think once somebody makes a start (finds out a way to do it) it won't be too long before we can have a 3d library that uses (limited) 3d acceleration
As for the PSX emulator. I'm not sure if the PSX has hardware accelerations, but my guess is that they just emulate that hardware chip, so basically it's software rendering.

 

[Radek]

Unless loading from a CD is done by a BIOS call (where HLE would be necessary), the two should be exactly the same. I don't know much about the PSX's internals but that's how it is in other systems.

They are the same in one sense but selfmodifying code is doing its "thing" (selfmodifying itself) rather on a regular basis (even milions times per second) where reloading code from CD is done rather at bigger time intervals.

If you mean is there a hardware accelerator in the GP2X then: no
I've seen discussions about the possibilty to use the coprocessor or the GPU to accelerate floating point operations which could be considered hardware 3d acceleration. Some people believe it's possible but wheter it's possible or not, it's very difficult I think.
The real hardware guru's would have to take a look if it's possible. I think once somebody makes a start (finds out a way to do it) it won't be too long before we can have a 3d library that uses (limited) 3d acceleration
As for the PSX emulator. I'm not sure if the PSX has hardware accelerations, but my guess is that they just emulate that hardware chip, so basically it's software rendering.

Yes - PSX has a hardware 3D accelerator. It's not exactly advanced (it can not do a perspective correct texturing or z-buffering) and thankfully so. Therefore It can be done in software at decent perfomance.

 

[STTrife]

Indeed, I was more talking about code that change SOME of the existing code in order to do nasty tricks (for security, speed or something else)
Code that replaces large blocks of code with new code loaded from CD is somewhat diffrent, because I think that most game use exactly the same method of doing this (or not?) like something that the compiler of psx programs does for you. And not some nasty trick a ASM programmer built in.

 

[Radek]

Indeed, I was more talking about code that change SOME of the existing code in order to do nasty tricks (for security, speed or something else)
Code that replaces large blocks of code with new code loaded from CD is somewhat diffrent, because I think that most game use exactly the same method of doing this (or not?) like something that the compiler of psx programs does for you. And not some nasty trick a ASM programmer built in.

I guess there will be four options:
- normal interpretor (for best compatibility)
- dynrec with checks (for a compromise between perfomance and compatibility)
- dynrec without checks (for the best perfomance)
- dynrec with hand made patches to allow some nasty games to be run anyway under dynrec (will need extra work just to run one title that way)

 

[STTrife]

I read something about speedhacks too. It was a trick to improve the speed of 1 game very much (so it has to be done for each game, if possible). The thing was that developers made loops to slow down each frame until they got the right framerate, which caused a lot of unneseary code. The would alter the ROM and remove that loop (or jump out of it) and that would increase that gamespeed very much. (at least that's how I understood it worked)
However I have no idea if PSX games use such stupid loops anyway.
But maybe there are other speed-tricks with altering roms, and maybe that means I can play FF7 at full speed, which I really want

 

[mtracy]

Does anybody know when the next psx emu is going to be released? I thought it should have been released yesterday at about 6pm. :huh:

 

[STTrife]

No I think zodttd only said he was going to release more info on his progress.
And he said something about POSSIBLY a week (at least) before he would get it working.

 

[Knoxximus]

Yup.....progress report soon.....possible release near the end of the week. B)

 

[Nr.56]

There are no more graphical bugs in soul reaver. it looks very nice.
The game speed is unchanged. the bug before the first fight too.

12 of 20 games of my psx-collection function on this emulator.
a question : is that below average?

Wipeout 2097 is playable at the offered 9-11fps. i think it is not necessary to get 50 fps to have fun with 3d-games. most of 3d-games impress at 20 frames. if the recompiler is able to achieve that (in case of wipeout it means 2x faster), everybody should be pleased because testing changes to gaming!

 

[Foton]

Maybe a stupid question, but is PSX4GP2X use both processors now?

Would it also be possible to play the music (audio tracks) like in the Wipeout series?

 

[STTrife]

As far as I know they don't use the coprocessor now. I believe there are some limits like they can't run asynchrone so apperantly it's hard to use them both...
Someone can explain in more detail???

 

[Shikaku]

As far as I know they don't use the coprocessor now. I believe there are some limits like they can't run asynchrone so apperantly it's hard to use them both...
Someone can explain in more detail???

The processors fights over memory if both are used with memory, so they would either go out of sync or crash. It has been proposed that a good way to use the second processor is using a 4K sound instruction/decoder, but that has to be done in ASSEMBLY and be very compact (meaning hard as fuck basically).

 

[STTrife]

So what is the coprocessor good for at this point...? nothing?

 

[greenmikey]

Its safe to say coprocessing..heh. Other then eating memory..eh

 

[Shikaku]

So what is the coprocessor good for at this point...? nothing?

I said hard as fuck, not impossible. If someone were to do it, it would be almost like running an emulator at a speed without sound.... with sound

=)

 

[zodttd]

Hi,

Quick progress report. I've been working non-stop on the dynarec. I have the BIOS dynamically recompiled. There's a weird compiler issue I'm sorting out right now and once that's done I should be able to get games recompiled as well.

So since the BIOS is recompiling at this point, and not games, it's hard to tell what speed increase is noticed. My guess at game performance with the dynarec still stands at 2x to 3x FPS.

Unai, sorry I have been out of touch, I've been working on this dynarec all this time.

A special thanks to those who donated for my work on the dynarec!

Hopefully next progress report will include recompiled games working. ;P