Search results

got the speed back up again but compatibility problem is still there. I think there is a problem with c_out. I just deleted it cause I thought it was useless. I now think that it matters when doing rotations otherwise the c_flag is not set properly. I'm going to use devkitgp2x on windows from...

It runs on the gp2x if that's what you mean, but I' afraid I messed things up further so now its compatibility is very bad and also slower than before.

I managed to get vba2x to compile with devkitgp2x. If anyone is interested, I can e-mail the src to you. I'm quite busy over the next two months so I hope someone else can take over this.(Also I'm out of ideas)

I think Slaanesh and Franxis makes a good team. Maybe they can port GP32 and GP2X versions together

what about q3? was it the same?

did it work before? which version? Did it work in previous versions?

http://www.gp32x.de/board/index.php?showt...=32899&st=0 - This looks most promising. It has a dynamic recompiler and FRANXIS may work on it. The main changes as compared to the original vba: 1) removed filter code. (The ones that make the screen looks smooth when you scale up on a pc) 2)...

q.4 uploaded. please test. I need to know which asm opcode is giving the problem.

#define OP_SBCS \ {\ asm("mov r0,%3,lsl#29\n"\ set bit 29 in R0 based on C_FLAG "msr cpsr,r0\n"\ copy into the status register "sbcs %0,%5,%6\n"\ subtract with carry "mrs r0,cpsr\n"\ copy status register into R0 "mov r1,r0,lsr#30\n"\ extract Z_FLAG "and %1,r1,#1\n"\ "mov...

Hi anyone tried porting this ? http://dl.qj.net/gpSP-v0.8-1.0-1.50-Eboots...10546/catid/122

Thanks for testing . What about the other games? If they work, it means the problem is isolated to the rotate and shift part. It will be easier to check through those only.

please try version q.3

I've uploaded version q.3 Please check. I've put back the c-core for the shift and rotate instructions.

I'm not confident about these as well. #define LOGICAL_LSL_REG \ {\ u32 v = reg[opcode & 0x0f].I;\ C_OUT = (v >> (32 - shift)) & 1 ? true : false;\ value = v << shift;\ } #define LOGICAL_LSR_REG \ {\ u32 v = reg[opcode & 0x0f].I;\ C_OUT = (v >> (shift - 1)) & 1...

Sound to 2nd CPU has been done by notaz for picodrive. I guess the technical mastermind devs would probably find it more challenging to do gfx on 2nd cpu. Since notaz could probably do the sound if he wanted and had the time to do so.

can anyone try with sound off from the cfg file so I can see if its the sound part which is causing the problem?

great. I think you are right unless you do fullscreen stretching in c. with hardware, this is done without any appreciable effort. I guess I can stop now cause I seem to be causing more problems with the recent changes than I am solving.

If its from so far back, then its probably not the arm asm cause that's starting from n.

Thanks for the encouragement I have learnt a lot from doing this. I think it would have been a lot easier for a someone who already knows asm. As far as I can see now, at frame skip zero, cpu emulation is taking about 20% of the time. About 70% of the time is on graphics and the rest probably...

Thanks. Can you try with q.1? did they work with p or o?(although that probably doesnt matter so much as I probably wont remember what I changed there)