Actually, it has. Some long-standing bugs have been fixed. Vertically oriented games now work, snes emulation is better, there's a longer list of support games and other fixes in fba2012. There's a limit to what you can get out of a system, no matter how much you optimize it. The people hoping to get 25% increses in fps on pcsx or the like are delusional. The performance of pcsx-rearmed, in general, not just on the 3ds, hasn't improved in years, and most likely it won't improve much as the hacks that would be needed for that would certainly introduce regressions.
People say stuff like "but even the original DS had snes emulators!!!", and to that one can say "yes, it did, but those sucked, most games wouldn't play, and the ones that would were plagued with all sorts of bugs". That's because of the hacky nature of those emulators, which was needed to even get the stuff to work on that system. Most devs nowadays don't want to waste time on hacky codebases, but rather work on an actually good emulator that has a chance of being largely compatible with the system's library. If that means, for the majority of people, that the 3ds doesn't have 'good' homebrew emulators for consoles more powerful than the genesis, well, tough luck.
I don't understand exactly what you mean thee, but it's quite clear that a 240p sceen should be using the RGUI interface.
Umm.... the 3ds has native support for gba games. You can't upgrade the cpu/gpu of a Wii, or a ps3, or a raspberry pi, so the second statement is strange to me. The fact that the crashes are apparently random (i've never had the issue as i've never used that core) also makes it rather hard to try and debug. Of course, it's not impossible, and you could try to narrow down the source of the problem by having a look at the code, or attempting to compile the software with no optimization flags and stuff like that. That's what I would do if I cared about that, but since I don't I dedicate some of my valuable spare time to fixing the bugs that actually bug me.
You don't need to organize that with anyone, or ask permission to do so, just open the bounty and put some money towards it. If others care to chip in and the bag gets large enough, someone may be compelled to try and figure it out. The amount of research you'd have to do to get into the ARMv7 (w/thumb) dynamic recompiler code is probably a major effort, though, and for the bag to get 'large enough' it would require to be of similar magnitude, or someone with just the right background who would enjoy a challenge with some beer money at the end.
