Search results

I assume it would be a bit better if the RAM clock was increased to 200MHz.  At 166Mhz (same as 3530) there wasn't much improvement.  I wasn't able to overclock much above 1.2GHz, but the failure mode was a bit unusual.  At 1.3 GHz, the core kept running but the I/O became unstable.  For example...

When playing Saturn games, I usually map A, B, C to A, X, B, and the second row of buttons to the keyboard.  The second row of buttons isn't used much in most games. I find the 4 buttons on the pandora to be a little too close to the right edge, and I have to bend my thumb to press the B...

For N64, I had to exclude code located in the first 4K of RAM to avoid some bootcode where the normal stack was not set up yet, but otherwise most games worked. I figured you'd have to exclude some of the BIOS code, but I guess it's more complicated than that...

What, no R29 hack?  Speeds things up a few percent and usually doesn't break anything.  It might crash if games use the ram mirrors though.

Oh, so the pointers are coming from clean_blocks().  I'll bet that can happen in mupen64plus too... There's a patch to relocate 0x80000000 in the 20110128 patchset.  It runs a few percent slower if you do this.  Notaz also has a mapping at 0x1f800000; I don't know if it's possible to relocate...

I actually fixed that in yabause (see this patch).  It would be possible to do it for MIPS emulation too.  It wasn't a huge improvement, and only affected a small number of cases, which is why I haven't put a lot of effort into it. Yes, that's a problem.  I suppose it would be possible to keep...

I looked at what Notaz did here..  basically the way I wrote it originally (for N64) it checks if the memory address is between 0x80000000 and 0x80800000.   If so then it accesses the memory directly (which is fast) otherwise it looks up a function pointer and calls a function (which is slow...

I recently got a Beagleboard-XM, and did a few tests to compare the performance of the DM3730 with the older OMAP3530, by running benchmarks of the mupen64plus dynarec on each. I expected that with the CPUs and RAM clocked at the same speed (500MHz/166MHz) that the DM3730 would be slightly...

Hmm... Is there anything else driven from the L3 clock besides the SDRAM, NAND, and SGX?  Increasing the clock speed of the NAND controller should still be within spec.  I don't know what would happen with the SGX though.

As far as I can tell, DPLL3 (CORE_CLK) works the same way, and the RAM is clocked at half this speed (L3_ICLK).  It's controlled by bits 16-26 of CM_CLKSEL1_PLL (0x48004D40). Setting this is a little tricky, since it apparently needs to be done while the RAM is not being accessed.  Normally it...

I haven't tried it.  Most of the clock frequencies are actually set up in the uboot before the kernel loads.  The DM3730 is speced for 200MHz RAM (instead of 166MHz) but I'm not sure what would happen if you tried to use those timings with OMAP3530.

It would be possible to benchmark the beagleboard-xm against the original beagleboard, if someone wanted to buy one of each.

In addition to overclocking the CPU, you could overclock the RAM.

I thought he meant replacing something like mov memory,%ax add $1,%ax mov %ax,memory with movzwl memory,%eax add $1,%eax mov %ax,memory or ldrh r0, [r1] add r0, r0, #1 strh r0, [r1] To do that, you need to determine that nothing will read the upper half of %eax. Then the...

I did this for the 64-bit registers in mupen64plus. Basically it does a seperate liveness analysis on each half of the register. If it determines that only the lower 32 bits are used, then it can discard the rest. The same can be done for 8 and 16 bit registers.

@@ -2929,6 +2941,19 @@ inline_writestub(int type, int i, u_int addr, signed char regmap[], int target, } //emit_pusha(); save_regs(reglist); +#ifndef PCSX + // rearmed note: load_all_consts prevents BIOS boot, some bug? Well, yes... This code is supposed to handle a pagefault. It...

I bought this one: http://www.amazon.com/gp/product/B0007V46HK

In the dynamic recompiler that I did for yabause, I often save and restore the carry flag with adc/lsrs. This does set the N and Z flags (and adc also sets V). If you want to set the carry flag, you can use: cmp reg, #0Similarly, cmn #0 will clear the carry flag (and V flag). This will...

http://wiki.yabause.org/index.php5?title=SH2_dynamic_recompiler

It's built with GTK. It should be possible to build with Qt though. Just make sure that the dynarec is selectable in the GUI. I told the Yabause developers what I was working on.