The only way for plenty of 100% emulation is....


RoboJoe32

Active Member
Joined
Mar 9, 2003
Messages
815
Most emulators run slow on GP32, not because of hardware limitations, but simply because the software for the ported emulator is NOT specificly coded for GP32 hardware (ala NES emu for GBA works because it was CODED for it, not ported) So if there are any skilled programmers out there, make your emu from scratch (or enlist the help of the greatest emu coders of them all (ZSNES team, NESTICLE team, WGens team, etc.) for assistance on the project.

So if gp32 is to ever hope to have 100% speed SNES, NES, genesis, GBC, and more emulation (maybe even neo geo) Instead of porting an existing emu, Make one that uses the GP32 hardware to it's maximum and is coded around it. Anyone else agree?
 
true.

but a port is easier to do (I'm certain!) and definitely preferable to .. nothing!

if the scene were bigger (more people owned a GP32 - hell, more people even knew what it was!) you can bet there'd be more support for it in the development community, and people would be more interested in picking it apart and writing great apps for it. While it's true that there ARE some great GP32-specific programs, it's obvious that the dev scene's not so big as, say, the GBA scene (not terribly large itself!) or even the older GB was in it's day.

All we can do is support the system and hope it gets more recognition and exposure. It's quite powerful and capable, and if more folks know about then it will get the attention it deserves. Until then it's just gonna have more ports than system-specific software, like say a Pocketnes.

(again, let me reiterate that ports are still pretty damn well done, all things considered! It's never gonna be as good as a specific one though, everyone knows this - but you gotta take what you can get and be grateful for it!)
 
Its all well and good pointing out this but the time and effort it takes to do this is a different matter.

Also -- I think virtually all gp32 programmers will know this already!!!? - after all their programmers and this has been pointed out before.

If only the community is full of skilled programmers with a willingness to such ports - there is a couple out there who have done so well with the ports already which I am more than greatful for!!

--but why bother doing an emulator from scratch if a port can suffice with full speed and sound(eventually!)

--Thank you to all those people out there who have done so much work for this community already! and for everyone involved in any shape or form! (I sound like a bit bum licky)

--and even if any production\programming halted right now with the gp32 - I am very happy!! anything more is a bonus!!
 
Well, the emulator that surprised me most is the PC-Engine-one.
Full speed with full sound, some minor slowdowns (but hey - it's the first beta!)

All other emulators work fine - until you turn the sound on. Then they skip and get slow.
AFAIK the PC-Engine is the only hardcoded one...

But I'm pretty sure a GP32 CAN emulate a NES or a GameBoy Color properly with full sound and speed.

Ah well, I forgot the SMS Emulator, which runs hell fast, too. Is that hardcoded?
And Frodo works quite nice... but that's ported (I guess a C64 doesn't need that much computing power as a PC Engine - after all, the first emulators around WERE C64 ones... 1991 or so :)

But you're right: Better than nothing. And I REALLY appreciate what all programmers here are doing just for the enjoyment of the people :)
It's just good to know that fast emulators actually ARE possible.

And when I take a look at ports of emulators for other machines: The longer they exist, the faster they get. Because a programmer changes the code time after time, and after a while there's not much of the port left, only GP32 specific code :) So, just be patient :)
 
As to GP32 "evangalism", whenever there is a GBA article on slashdot, I add comments comparing the GBA to the GP32, with links to here, gbax.com, lik-sang, ect.

There are large numbers of hacker types on slashdot and other sites that we need to convince. The more people working the more good stuff is available. The more good stuff is available, the more people buy GP32s and create more good stuff.

Working GBA and/or Linux ports would get full article coverage on several sites and lead to more developers.
 
I have to say that if Mike hadn't done the port of Frodo to GP32 then I would personally not be looking at mine on my desk. I brought my GP32 solely with the intent of playing my favourite old C64 games on it, and couldn't be happier. Full speed and with a fairly decent audio emulation, what more can you ask for. Sure some of the other emu's are not upto scratch yet, but the publicity alone which the GP32 is now receiving
has to be in the major part attributed to these "ports" of C++ coded emus.

Of course I have since come to realise that there are other great programes (aside from Frodo) for my GP32, I am a big BIG fan of the PCEngine emu, and SMS32, both of which demonstrate the capabilities of this hand-held wonder of modern science!

PK.
 
Hardcoded emus are very very nice... but they are a LOT harder (quote from Rattboi, I think "It's a port - I'm not insane" regarding Wonderboi 32).
Add to that that some emus ported will eventually end up full speed anyway - a la Snes9x (with a bit of luck)... Since its all being re-done as ASM and similar, speed can be expected to steadily increase as the original source becomes more and more of a framework and less and less of the actual body of code... Albeit it isn't there yet, but compare 0.8b to 0.32a and you'll see what I mean... the speed difference is **massive**.

Oh, and BTW: Zadoz's Snesemu is a hardcoded one as well... really waiting with baited breath for the eventual next releas of that one :)
 
As Tobriand said, SNESemu is hardcoded and very fast. So I think it is important to have hardcoded emulators, as speed is a big issue with the GP32, it can only emu such hardware as the SNES or GBA effectively that way. However - the easiest way is to port it, and that's also important. For example a lot of people are dying to play GBA so a port of BoycottAdvance or similar would be relatively quick and easy compared to hardcoding. Then people can begin to optimise and rewrite parts of the port so it gets to a sensible speed.

- Rico
 
A technique which is not very well known but works very nicely on hardware like
this which is *almost* fast enough is static binary translation. My tailgunner port
for the GP32 is a translation of cinematronics machine code to low level C. I'm currently
experimenting with the same technique and a Z80 translator in order to port a
version of MS Pacman to the GP32. Although Skeezix's emulator plays it OK (I
think he had to crank up the speed to 133MHz though), it doesn't have enough steam
to rotate the screen sideways and stretch the image to full screen size. I'm fairly
sure that a static translation will be fast enough that I can scale the image on the fly.
(Tailgunner runs at full speed at the default 66Mhz and highly non-optimised graphics...)

Binary Translation HOWTO: http://www.gtoal.com/sbt/

Current 'in progress' snapshot of Z80 -> C translator: http://www.gtoal.com/sbt/z80/
(z80.c is the translator; file mspac.z80 is the current output of the translator - this is 2 days
of work so far, so it's not ready to run yet, but probably not much more than another
weekend's hacking away...)

Graham
 
Back
Top