| Main > Main Forum |
| PiMame- are we finally there? |
| << < (8/13) > >> |
| shponglefan:
Maybe it's just me, but I don't think development on limited platforms is necessarily a waste of time. In fact, optimizing code for limited platforms can be a greating learning process. And (hopefully) these lessons learned will carryforward to next generations, whether the Pi or something else. |
| Haze:
--- Quote from: shponglefan on September 22, 2013, 07:33:48 pm ---Maybe it's just me, but I don't think development on limited platforms is necessarily a waste of time. In fact, optimizing code for limited platforms can be a greating learning process. And (hopefully) these lessons learned will carryforward to next generations, whether the Pi or something else. --- End quote --- Learning what tho? Most phones you buy these days have more power than a desktop from a couple of years ago; optimizing for really low spec systems is becoming a specialist area with more limited real world applications each passing day. Finding people to do real emulation work these days is tricky enough, we're seeing more and more who are only interested doing something if there is big money in it and spending years training your skills to write optimized code for a specific low-spec platform / CPU and the quirks of however it's hooked up isn't really one of those things with technology changing so quickly and diminishing returns for your efforts (even if it is rewarding when you do) The "fastest code" way to emulate anything / optimize an emulator is to not emulate it properly, take shortcuts knowing the games you're targetting don't need specific codepaths, or won't make bad accesses you have to spend extra code safety checking etc... is that a good lesson to be learning in a world demanding *secure* code with all bases covered? Sadly a lot of the cases where learning such skills would be useful (actual viable commercial platforms) end up being locked out from regular people, although if somebody was to want to learn such skills a hacked up console and some stolen devkits would still be a more relevant area than a Pi... The problem with such skills is that you're effectively starting at 0 with each new platform because techniques that work well with one CPU / platform might not work at all well with another. |
| Typefighter01:
Guys like Shea are "probably" laying the emulation support groundwork for future verions of both the Pi and Linux that will certainly be released with more power and memory, a Model C :dunno . It might not be the current version of Pi that delivers the goods, but it will come and we will all reap the rewards. Haze-I bet you and a few of the MAME devs could modify a version of MAME to work nicely with the Pi...just kidding :cheers: |
| shponglefan:
--- Quote from: Haze on September 22, 2013, 07:40:52 pm --- --- Quote from: shponglefan on September 22, 2013, 07:33:48 pm ---Maybe it's just me, but I don't think development on limited platforms is necessarily a waste of time. In fact, optimizing code for limited platforms can be a greating learning process. And (hopefully) these lessons learned will carryforward to next generations, whether the Pi or something else. --- End quote --- Learning what tho? Most phones you buy these days have more power than a desktop from a couple of years ago; optimizing for really low spec systems is becoming a specialist area with more limited real world applications each passing day. --- End quote --- Except we aren't talking about $500+ dollar smartphones; we're talking about a $35 computer. Given the success of the Pi, there is clearly demand for such inexpensive systems. And while we will see an increase in power with successive generations, the upper limit of emulation will keep getting raised. Heck, we only now have desktops barely powerful enough to run emulated Blitz at a decent speed. --- Quote ---The problem with such skills is that you're effectively starting at 0 with each new platform because techniques that work well with one CPU / platform might not work at all well with another. --- End quote --- Maybe, maybe not. We're not yet sure what a next-gen Pi will look like. |
| paigeoliver:
--- Quote from: Typefighter01 on September 22, 2013, 07:55:13 pm ---Guys like Shea are "probably" laying the emulation support groundwork for future verions of both the Pi and Linux that will certainly be released with more power and memory, a Model C :dunno . It might not be the current version of Pi that delivers the goods, but it will come and we will all reap the rewards. Haze-I bet you and a few of the MAME devs could modify a version of MAME to work nicely with the Pi...just kidding :cheers: --- End quote --- I don't know if there is any point of laying a ground work based on tweaking the code to mame .17 or whatever version from the 233 mhz processor era that pi mame is based on. From a game playing perspective mame didn't get GOOD until the .50s releases, which hit the hardware a lot harder than .17. Then you have the other Pi problem, it doesn't output video in a format that is really friendly to arcade game emulation. |
| Navigation |
| Message Index |
| Next page |
| Previous page |