Dolphin, the GameCube and Wii emulator - Forums

I don't know how to compile those three files. Does it solve the stuttering problem?

Those three "files" are three lines of code that were changed in that revision, specifically revision 3.0-685. Just use builds off the main site. The stuttering starts to occur between 3.0-700 and 3.0-710, and only when using DSP LLE. So to avoid stuttering with LLE, just use 3.0-700 or earlier. Unfortunately there are still going to be some issues though; the devs are looking into it.

Personally I've had alot of success with 3.0-600 and Prime 1, but I haven't done enough testing with Prime 3 to see if 3.0-600 fixes the BGM looping issue it has. I need to do that sometime.

i had major problems with the music glitching and rapidly looping in the PAL version of the game using HLE with 7.035. it seemed to arise in conjunction with a few other glitches- once the music glitched, no speech would play during cutscenes even though the sound effects were working perfectly, and often i couldn't see "press A to talk" when near a NPC. sometimes, i could see the message, but pressing A did nothing.

some cutscene events refused to happen while the music glitch was occuring, so the game was practically frozen at these points. one of these moments was in a bomb slot at the start of the game (same place chenci posted)- i could lay as many bombs as i wanted but the switch would never activate, and it wouldn't release me until it had been activated, so i was trapped inside it.

another was the first time using the command visor to land samus' ship on norion- once the visor was used, the cutscene showing samus' ship landing began, but the ship never appeared. the camera was just stuck on the sky waiting for the ship, and i had to restart from the last save to get out of it.


i took the advice above and downloaded a version before 3.0684 (3.0-655) and it's working great now, the audio isn't fucking up and it's even running at a better framerate than before. this game looks by far better than the first two, i'm kind of surprised by just how much more it's gained from dolphin. it doesn't work at all with recent revisions, though.

for what it's worth i found d3d11 to be the fastest backend (when running at 3x IR with default settings other than accurate vbeam emulation and openmp). with framelimit off d3d9 got 77-78fps, d3d11 got 82-83, and open GL got 75-76.

(08-19-2012, 04:04 PM)ironman120693 Wrote: [ -> ]Oh and also, since I leave frameskipping on 1(sometimes now) the framerate can vary from 20-50 and still be full speed.

So frameskipping on stage 1 makes speed independent from fps..?

I managed to play MP2 with EFB to RAM, with the sheer force of my cpu..runs fine @fullspeed most of the time, but in the command center the fps drop to 30 o_O..

i didnt test it yet, but am I able to play MP3, too, when im able to play MP2?

Then I have two questions left: DefenderX recommended to try rev 3.0, but my wiimote doesnt connect with this rev. "Only Microsoft Bluetooth Stack is allowed". Fuck that.. So i wasnt able to test it with 3.0, and Im running the newest rev now. There any way I can connect without installing Microsoft B Stack (which will probably not work <_<)

3rd question, I played MP2 with fullscreen and in window mode, and I got the strong suspicion that it runs better in window mode. Thats normal and due to the lower resolution of window mode, right?
Im not sure at all, because you cant see the speed at fullscreen, only the fps, but I think window really works better.

(09-02-2012, 06:26 AM)ESChinski Wrote: [ -> ]

(08-19-2012, 04:04 PM)ironman120693 Wrote: [ -> ]Oh and also, since I leave frameskipping on 1(sometimes now) the framerate can vary from 20-50 and still be full speed.

So frameskipping on stage 1 makes speed independent from fps..?
....
....
3rd question, I played MP2 with fullscreen and in window mode, and I got the strong suspicion that it runs better in window mode. Thats normal and due to the lower resolution of window mode, right?
Im not sure at all, because you cant see the speed at fullscreen, only the fps, but I think window really works better.

I honestly don't know. I keep testing, hoping to find definitive answers, but keep seeing contradictions. For example:

Metroid Prime 1 plays at 60fps for full speed. Well, you only need VPS to be 60 to get the gameplay to be full speed(sometimes this is inaccurate and the VPS is really equal to the FPS, this happens in Prime 3 for me) so frameskipping should make FPS independent of VPS all the time. The question should be the other way around, if frameskipping is OFF is FPS independent of VPS(gamespeed)?

I tested this on Metroid Prime 1 on both GC and Wii versions and the game is the same with or without frameskipping, as if frameskipping is already built in to the game itself. I don't know why I experience this, but it makes me mad because the VPS and gamespeed readings are inaccurate on Prime 2 and 3. I'll often see 100% for both games, when I'm actually not even playing full speed, you can tell even if you're not in windowed mode, if the game is full speed or not. So when I enable frameskipping, it makes those inaccurate readings accurate. Its driving me nuts! Maybe its just a bug, idk, but I always thought that without frameskipping, the FPS should always be the same as the VPS. But then I read that some games play at 30 fps, but VPS must be at 60 fps to be full speed, so some games will have FPS always half of VPS. So to answer your question... I just don't know.

To answer the other question, if you're using an AUTO resolution option, then I think yes, windowed will always outperform full screen, unless, your window is the same resolution or lower than what you're using for the full screen resolution(doesn't always have to be your monitors max res). Or if you use a locked resolutions(1x, 2x, 3x, etc.) the performance shouldn't change between windowed and full screen.

Windows mode outperforms fullscreen mode on "auto" because it lowers the internal resolution to fit the window size. At windowed mode it's 1x (usually anyway) and fullscreen it's 2x-3x.

Quote:I managed to play MP2 with EFB to RAM, with the sheer force of my cpu..runs fine @fullspeed most of the time, but in the command center the fps drop to 30 o_O..

Revision and settings? The prime games are very reliant on the revision you are running, thanks to some bugs and such. If you really want EFB to Ram speed (and don't mind using HLE cause of the LLE stuttering), go hashless. It's EFB to Ram speed is awesome.

(09-04-2012, 05:47 PM)MaJoR Wrote: [ -> ]...go hashless...

What does this mean?

http://dolphin-emu.org/download/branches/

Choose a build from the hashless branch.

(09-04-2012, 06:37 PM)DefenderX Wrote: [ -> ]http://dolphin-emu.org/download/branches/

Choose a build from the hashless branch.

what does hashless mean? you can explain it in german maybe

well, I got the impression that frameskipping 1 weakens the performance peaks, so at spots where the fps falls to 30 (command center) the game still runs fullspeed, because speed is still 100%, in opposite to frameskip 0, where speed at 30 fps is 50%.
but i got also the impression that frameskip 1 is a bit worse for spots which are not demanding, theres frameskip 0 better I think.

(09-07-2012, 03:18 AM)ESChinski Wrote: [ -> ]what does hashless mean? you can explain it in german maybe

Speed of EFB to Texture + Accuracy of EFB to RAM