DraStic

I recently had reason to go on a trip and so while waiting at the airport and while flying I fired up some of my emulators to finish off some partial games. That is when I noticed something interesting...

DraStic on my Galaxy Tab S uaes more power per hour then PPSSPP, or epsxe, or fpse, or reicast, or some cores in retroarch (though I only did some testing in retroarch and fpse after the trip). I was kinda curious as to why this may be though I don't know if there is a reason.


Btw I also realized how spoiled Exo and Lordus have made me. Exo made an emulator that works pretty much 99% of the time with almost no settings needing to be changed between games. Unlike PPSSPP, though it has per game settings. Lordus has made a UI that works (unlike retroarch), makes sense, is fast to navigate, and the quick save and load feature that does not requiee to go all the way to a menu to use. So thank you for helping me stay entertained while traveling and curse you for setting the bar so high for other emulators I keep noticing all the things they don't quite do as nice as DraStic instead of just playing my games.

I don't know for sure but the principle reasons are probably:

1) PPSSPP renders the screen using OpenGL ES, DraStic renders it in software and has to do the 2D + 3D parts separately. There are a variety of reasons why it's easier to do hardware rendering with something like PPSSPP, but the GPU is just much more power efficient than the CPU at this task. This is much more the case if you have high res rendering on. It takes a lot of grunt from multiple cores to really pull it off.

2) PPSSPP uses a lot of high level emulation at the OS/SDK level and not the hardware level. This works with PPSSPP because the PSP uses a fixed firmware to implement these software functions and all games utilize it, and the games don't really have direct access to a lot of hardware functions. This means lower overhead emulation and that they may be able to make some simplifying assumptions for CPU emulation.

I'm still looking at the possibility of OGL ES rendering as an optional alternative, although I don't really know how well it'll work since it requires some very tricky handling of 2D + 3D integration.

I didn't realize that the gpu was more power efficient. I guess the processing speed advantage not only can increase certain tasks performance, but the performance advantage would also lower power usage through less computational runtime needed for the task.

Strange. While with uprendering and xBR on I could iron a shirt, with those off and every other frame skipped I could play for at least a hour with ~20% battery loss on my Nexus 4. Given the conditions, this is very battery-friendly. I'm mostly playing Pokémon, which might not be the most resource-hungry, but certainly isn't the least.

Is sound latency also a big waster? I recently dropped it from low to medium to reduce cracking, but aside from that there's no difference. How many miliseconds are each step worth?

Fefo wrote:Strange. While with uprendering and xBR on I could iron a shirt, with those off and every other frame skipped I could play for at least a hour with ~20% battery loss on my Nexus 4. Given the conditions, this is very battery-friendly. I'm mostly playing Pokémon, which might not be the most resource-hungry, but certainly isn't the least.

Is sound latency also a big waster? I recently dropped it from low to medium to reduce cracking, but aside from that there's no difference. How many miliseconds are each step worth?

Iron a shirt?

ericbazinga wrote: Iron a shirt?

Based on the heat produced.

Fefo wrote:Strange. While with uprendering and xBR on I could iron a shirt, with those off and every other frame skipped I could play for at least a hour with ~20% battery loss on my Nexus 4. Given the conditions, this is very battery-friendly. I'm mostly playing Pokémon, which might not be the most resource-hungry, but certainly isn't the least.

Is sound latency also a big waster? I recently dropped it from low to medium to reduce cracking, but aside from that there's no difference. How many miliseconds are each step worth?

While there are differences in batteries and processors my galaxy tab s 8.4 was using about 20% per hour (so not a big deal for a 3-4hour flight, not to mention my power bank), but that was with my usual setup. 0 frameskip, multi-threaded off, sabr filter w/ lcd3x.... so not really power saving settings. To be fair I do tend to play most ppsspp games at 3x internal resolution and native 1600x2560 display resolution, but settings do tend to vary much more wildly then with DraStic (I was playing valkyria chronicles 3 at 3x internal and native display with 1 frame skip auto since the 2d visual novel parts need it for some odd reason but not the 3d gameplay).

3x... that's brutal even for my PC.