Dolphin, the GameCube and Wii emulator - Forums

No,it will not. You will get same framerate you usually get,it just that you wont be able to control the game,as far as I understand.
I dont really find this very useful,we are interested how games perform on DIFFERENT systems,we already know how they perform on our systems.
If we want movies,just ones recorded with fraps and uploaded to for example youtube I find more useful,as they will show how games perform on different systems

Zee, read what Shonumi said. He explained it twice already.

Rpglord, you do not understand the purpose of .dtm it seems.
Read what Shonumi said, and keep reading it until you understand.

Recording with FRAPS will kill your performance and is not a viable option for those with worse rigs.
Even so, one user could run around in Windfall Island (Wind Waker) and report his progress there while the other runs around in Outset Island (Wind Waker).
Different locations have different impact on the benchmark obviously.

Even if we agree to run around in one location, the actions we will do will never be the same which will give everyone a different result depending on what he did.
One could stand still for 2 mins on one location and get a better average framerate than you for example.

That's where a .dtm file comes in. Everyone gets to benchmark the same actions in the same environment to get the most accurate results as possible.

The .dtms are probably a moot point at any rate. The latest revisions still don't seem to fix the issues I was having. I still get the "Fatal desync. Aborting playback. (Error in PlayWiimote: 23 != 4, byte 130)" errors. Even when it does work, the input is still way out of sync when I replay it, or the input is completely incorrect (running forwards or crouching instead of jumping). This might have something to do with the speeds at which I'm trying to playback the movie. It seems if the playback speeds are significantly different from the speed the movie was recorded, the input gets desynced.

For now though, I've a couple of areas that demonstrate the gaps between HLE and LLE audio. Even with Accurate VBeam enabled, LLE is still behind at least 14% in my tests. Screenshots after work are coming. The areas are pretty clear where they are, and it'll only take people a minute of playing Yoshi Star to get into position.

So its somewhere in Yoshi Star..hm..ok I will test these areas too when I see you screenshoots
Garteal,thats why screenshoots work,so we are sure everyone is in same location doing same thing.
I agree fraps do make a perfomance impact..so I guess sticking to screenshoots is best option

Which dolphin should i use to test, also which settings, should i check Vbeam and uncheck Idle skipping?

You can test 3.0 using Starscream's speed settings. You can also use the VBeam Trick (VBeam enabled and/or disable Idle Skipping). Just be sure to do tests without the VBram Trick for comparison. I've tested the latest revision to see if the HLE vs LLE gap is more noticable now than before (and to see if delroth's improvement to the 64-bit JIT recompiler is any good :p) I just got home from work, gonna relax (hey, it's Friday, so...) then I'll post my results from earlier here.

EDIT: Results are up. Note, I compiled the latest rev myself, but the only reason it's a "dirty" build is due to the fact that my version of GCC doesn't support C++11 in some regards. I had to change "nullptr" to "NULL" on 4 occasions. I assume this really shouldn't have any noticeable bearing on performance.

Super Mario Galaxy 2 - HLE + VBeam - Yoshi Star : 1st Checkpoint
OpenGL Video Backend
Result: 99 FPS


Super Mario Galaxy 2 - LLE + VBeam - Yoshi Star : 1st Checkpoint
OpenGL Video Backend
Result: 78 FPS


Percentage Difference: +23.7%

Super Mario Galaxy 2 - HLE + VBeam - Yoshi Star : After 1st Checkpoint, by the sign
OpenGL Video Backend
Result: 79 FPS


Super Mario Galaxy 2 - LLE + VBeam - Yoshi Star : Right after 1st Checkpoint, by the sign
OpenGL Video Backend
Result: 67 FPS


Percentage Difference: +16.4%

I also came across a potentially interesting discovery that may explain quite a bit of things. Just need to do some further testing, think it out some more, and get some sleep. It's 3 AM atm...

Here are my results : definitely highest difference I saw so far in any wii game ( around 20% ) :
( Interesting that I got slightly higher fps then you here ) :




Looking forward to hearing about interesting discovery

My cousin has Resident Evil 0 GC EUR, I'll borrow him the disc for testing here, since i only have the wii version usa.

Dolphin 3.0 x64

[color=#32CD32]Resident Evil 4 Wii edition USA (RB4E08) on HLE + Vbeam and Idle skipping off[/color]

http://imageshack.us/a/img560/2337/72223586.jpg







http://imageshack.us/a/img694/359/70888299.jpg


[color=#FF0000]Resident Evil 4 Wii edition USA (RB4E08) on LLE with LLE on thread ON + Vbeam and Idle skipping off[/color]

http://imageshack.us/a/img607/6415/56398139.jpg







http://imageshack.us/a/img829/2203/50636159.jpg

[color=#00BFFF]Resident Evil 4 Wii edition USA (RB4E08) on LLE with LLE on thread OFF + Vbeam and Idle skipping off[/color]

http://imageshack.us/a/img715/5531/97734306.jpg







http://imageshack.us/a/img87/8127/56125604.jpg

I don't know why but when i'm using "LLE on thread ON" i get less FPS, with "LLE on thread OFF" i get more FPS, maybe because of the turbo boost o_O.

Hmm is there a way to "hide" the screenshot, something like "Show spoiler" and "Hide Spoiler" ?

Thank you for contributing, hyperspeed.
I dont see you getting less fps with lle on thread on,1 fps is in the margin of error.
But you should be getting more fps with lle on thread on.
Try enabling "lock threads to core" and see if that helps

(10-21-2012, 12:09 AM)rpglord Wrote: [ -> ]Thank you for contributing, hyperspeed.
I dont see you getting less fps with lle on thread on,1 fps is in the margin of error.
But you should be getting more fps with lle on thread on.
Try enabling "lock threads to core" and see if that helps

"lock threads to core" just uses more cpu here, same FPS with it ON or OFF.

[color=#483D8B]LLE on thread ON:[/color]





[color=#FFD700]LLE on THREAD OFF:[/color]








[color=#FF4500]Resident Evil Archives 0 (RBHE08) on HLE[/color]





http://imageshack.us/a/img90/7649/31973980.jpg





[color=#FFA500]Resident Evil Archives 0 (RBHE08) on LLE with LLE on THREAD OFF[/color]





http://imageshack.us/a/img717/6186/53173134.jpg



[color=#483D8B]Resident Evil Archives 0 (RBHE08) on LLE with LLE on THREAD ON[/color]






http://imageshack.us/a/img801/255/95395586.jpg