It's Gone A Bit Quite---gpx2

[Bobafeet]

Oct 11th: Dev section will open soon, GCC4 is now working -- it's now the 24th, soon we want it now!!!!! ........Gpx2 gone a little bit quite we want more video we want more infomation!!!!...........now........not soon!!!!....soon is what my momma used to say y'll get it soon son...never happend....

 

[TelcoLou]

It's GP2X ...

 

[DJWillis]

Spend some time looking at things like the WiKi and IRC (#GP2XDev on EfNet) and you will see it's far from quiet on the GP2X front . I should know, I am the one that got GCC4 to finally build executables that worked on the console .

 

[Squidge]

Yeah, I think most developers are bored with the gp2x now and are moving to other platforms... not

After spending ages trying to get alternative cores properly working in UAE, I'm writing my own 68k core in arm assembler, so it'll take some time, but should be very compatible.

 

[Reesy]

I like the sound of that, send me a copy when your done go on son you know it makes sense.

 

[Eolair]

Interesting...finding a progress report on the UAE project here. I like

 

[mr.mirko]

Yeah, I think most developers are bored with the gp2x now and are moving to other platforms... not

After spending ages trying to get alternative cores properly working in UAE, I'm writing my own 68k core in arm assembler, so it'll take some time, but should be very compatible.

​

Take a deep look to the 68k core.

And say this again...

btw, there are over 5000 cpu-opcode cases... But hey a skilled man can do it...
btw, the uae cpu core is VERY fast. It is sused in the hatari Atari ST project, and my 266Mhz Notebook can emulate hatari in full speed. So the main problem with UAE is the thing around the cpu core. Not the 68k core needs a rewritten, all other stuff around it needs it

 

[Squidge]

I know quite a lot about the 68K processor due to the fact I had an Amiga A500 followed by A1200 for many years and my favourite software was DevPak 2 So I don't think writing a new core will be that difficult (it isn't so far)

As for UAE, I agree. A lot of the graphics code need optimizing, but the CPU is not as fast as it could be either. For example, doing flag calculations in C for example takes up a LOT of time and code, and could be done a lot faster in arm assembler (eg. just move the cpsr into a register and flip the carry bit for most of it rather than loads of compare instructions). Then you can store the cpu context and execute multiple instructions instead of fetching it for every instruction. There's many other optimizations that can be done too.

Considering that a 68K core can be used on many other projects, I think it's more than worthwhile spending the time on it before attempting optimizations elsewhere (although I've already sectioned off pieces of code that should run well on the other processor providing even more speedups).

 

[efegea]

Squidge, there is already a 68k arm asm core. I think it's called cyclone. It's used on various GP32 emulators. Why not use it instead of rewrite one yourself?

 

[WoD]

Where's the challenge in that?

 

[Jarska333]

Squidge, there is already a 68k arm asm core. I think it's called cyclone. It's used on various GP32 emulators. Why not use it instead of rewrite one yourself?

​

Compatibility issues, I'd imagine. It didn't work in CaSTaway, atleast.

 

[woogal]

Squidge at there is already a 68k arm asm core. I think it's called cyclone. It's used on various GP32 emulators. Why not use it instead of rewrite one yourself?

​

After spending ages trying to get alternative cores properly working in UAE...

​