This pc runs dolphin at full speed

[Shonumi]

(06-19-2013, 03:43 AM)rpglord Wrote: Well I remember on one thread we were both posting screenshoots regarding some place which was gpu demanding ( or so we thought ) and we were both getting same results of well above 60 fps ( it was 70-80 fps I think )
Maybe something changed in latest revisions, I haven't tested it, but you could play it fullspeed on stock i5 around 3.5

Those spots I posted in were below 60 FPS for me; I distinctly remember them because I got slowdowns there regardless of IR and that's the only reason they were notable. I don't remember what your results were though. I tested 3.5 today, and it's still the same: fullspeed is unachievable for me, even OCed at 3.8 GHz. 3.0 is an entirely different case; I can easily run SMG 2 @ 3.8 GHz fullspeed for what I tested (not the whole game, obviously).

[rpglord]

(06-19-2013, 03:56 AM)Shonumi Wrote:

(06-19-2013, 03:43 AM)rpglord Wrote: Well I remember on one thread we were both posting screenshoots regarding some place which was gpu demanding ( or so we thought ) and we were both getting same results of well above 60 fps ( it was 70-80 fps I think )
Maybe something changed in latest revisions, I haven't tested it, but you could play it fullspeed on stock i5 around 3.5

Those spots I posted in were below 60 FPS for me; I distinctly remember them because I got slowdowns there regardless of IR and that's the only reason they were notable. I don't remember what your results were though. I tested 3.5 today, and it's still the same: fullspeed is unachievable for me, even OCed at 3.8 GHz. 3.0 is an entirely different case; I can easily run SMG 2 @ 3.8 GHz fullspeed for what I tested (not the whole game, obviously).

Well you made me go look for it, you are right you were not able to get fullspeed because of opengl which is clearly slower in this game :p
I got 66 fps with LLE and 72 fps with HLE, but it was on 3.0-787...
http://forums.dolphin-emu.org/Thread-lle-is-not-much-slower-in-wii-games-lle-benchmark-thread?page=13 ( post #123)

[Shonumi]

(06-19-2013, 04:08 AM)rpglord Wrote: Well you made me go look for it, you are right you were not able to get fullspeed because of opengl which is clearly slower in this game :p

In Dolphin, OpenGL performance will be determined by drivers and GPU make and model. There have been a handful of accounts comparing OpenGL on Linux and D3D9 on Windows; they should be on par for some GPUs (probably Nvidia mostly). The Linux drivers Nvidia provides aren't the same as their Windows ones; OpenGL performance is generally higher on Linux. Some speculate this is due to the fact that many games on Windows utilize Direct3D, so OpenGL gets marginalized on Windows, while OpenGL on Linux gets more attention given that it's the only API on that OS to implement. Once I get a working Windows 7 install, I'll test this out more thoroughly.

[rpglord]

(06-19-2013, 04:24 AM)Shonumi Wrote: In Dolphin, OpenGL performance will be determined by drivers and GPU make and model. There have been a handful of accounts comparing OpenGL on Linux and D3D9 on Windows; they should be on par for some GPUs (probably Nvidia mostly). The Linux drivers Nvidia provides aren't the same as their Windows ones; OpenGL performance is generally higher on Linux. Some speculate this is due to the fact that many games on Windows utilize Direct3D, so OpenGL gets marginalized on Windows, while OpenGL on Linux gets more attention given that it's the only API on that OS to implement. Once I get a working Windows 7 install, I'll test this out more thoroughly.

Yeah I know that, but we were using similar versions you used 3.0-782 and got 51 fps with LLE and I used 3.0-776 where I got 63 fps with LLE and 3.0-787 where I got 66 fps with LLE.
I think we used same settings, so I assumed it probably has to be something related to the OS/drivers/opengl

[Shonumi]

It's probably not the case that OpenGL was the culprit. If you look at the Wind Waker benchmark, you'll notice that your results are marginally better than mine (since you're a Sandy Bridge clocked 100 MHz higher than my 3.3 GHz stock when I added my results). Now the benchmark is intended for CPU performance, but remember we all used 3.0 with the same settings, with the exception that OS X and Linux users had to use OpenGL. Now if OpenGL were significantly slower in Linux than Windows, we'd see that affecting my results, especially with uncapped framelimits. But I'm not that far behind you. Now this doesn't mean it's true for all games in every situation, but it's some evidence of what I was talking about.

The reason for our performance difference is that Intel must have mislabeled your i5-2400 for some incredible chip from the future (that was backward compatible with your current mobo). Hey, it happens sometimes.

[rpglord]

(06-19-2013, 05:31 AM)Shonumi Wrote: Now this doesn't mean it's true for all games in every situation, but it's some evidence of what I was talking about.

This is probably explanation, SMG2 at that spot is exception where linux/opengl might be more then marginally slower.
This is not just the drivers support, dolphin code is also playing a role. There were some commits by rodolfo which provided a significant speedup in SMG2, but I think he only applied it to DX9/11, or atleast only when we were making the comparisions.

(06-19-2013, 05:31 AM)Shonumi Wrote: The reason for our performance difference is that Intel must have mislabeled your i5-2400 for some incredible chip from the future (that was backward compatible with your current mobo). Hey, it happens sometimes.

Well that could be the explanation too, but then why am I not getting that amazing future speed in WW too ?

[vbetts]

I can run Twilight Princess full speed on a dolphin. I'll have to get my video of the tv, gc, and solar pack I taped to a dolphin while I rode ontop another dolphin to play my game. Crazy times...