(06-17-2010, 07:13 AM)StripTheSoul Wrote: [ -> ]@ NeoBrain & Rodolfo (does he even ever visit this forum? 
): Ector said the stutters which mostly occur when camera angles are being changed, doors are opened, levels are loaded, etc. and which are veeery annoying, come from the slow shader compilation of D3D. Isn't there maybe a way to optimize the shader cache (which is already implemented in Dolphin) in a way that only new shaders have to be compiled? because it seems that every time all is being loaded again...for example when opening a door in Fatal Frame 4...
?
I don't think Tino ever visits the forums
And we are caching the shaders already, if you start the game a second time and go to the same place again and the stutters occur AGAIN it can't really be caused by shaders I guess...
hm...then it's not just the shaders, I guess...
congrats on the plugin and i hope it improves soon
Mario Kart Wii [NTSC] is quite playable in 5727, but there's some glitches, some new and some recurring... but nothing that affects gameplay:
â–º The old bug in the Miis' faces is still there;
â–º The (barely noticeable in many tracks) blue square taking up a quarter of the screen;
â–º This one is new: the FMVs get all screwed up after some time playing, even in the menus.
Apart from this, is very playable, congratulations to the devs for making such progress SO FAST. ;D
It is slower than DX9 for me, but I can almost tell that I saw some lighting effects with DX11 that were absent in the times I played it with DX9.
Ok, so I posted an FPS comparison of Rune Factory with DX9 and DX11 to my livestream account @
http://www.livestream.com/sibanell
You'll have to go to full screen mode with the video in order to see the FPS.
If this is useful / interesting to you, tell me and I can do something similar / more in depth with Final Fantasy Crystal Bearers.
Alright, any thought that the difference was just in my head is now definitely gone. I just filmed a comparison of the first 10 minutes or so of gameplay for FF Crystal Bearers, first using the DX11 plugin, then the DX9.
The difference between the two is tremendous, particularly the parts where the main character is battling the Yuke and piloting the Airship.
Additionally, both videos are almost identical in terms of the gameplay they display, but the DX11 video is about 9 minutes of gameplay (the first minute of the video is me messing up with Procaster ><), while the DX9 video is about 18 minutes long. Again, they show the same footage, the DX11 just renders everything nearly twice as fast.
Aside from switching between the DX9 and DX11 plugins, nothing else changes. I don't know if your plugin is just more efficiently written, DX11 is better for Dolphin than critics have forecasted it to be, or if your DX11 coding allows my video cards to take on a heavier processing burden, but whatever it is, the difference is already measurably clear.
NeoBrain, Xtreme2damax,
I have some old problem with fog calculations - check this issue please:
http://code.google.com/p/dolphin-emu/issues/detail?id=2589
and
http://forums.ngemu.com/dolphin-discussion/130629-resident-evil-4-issues-list.html
Besides - hack from EliotFur doesn't working. And I found the root of this problem:
real console hardware is capable of fog distance and range calculation, but Dolphin video plugins
still don't know about fog distance calculation - they are using fixed distance and simple linear fog
and this causes that issue #2589...
I can give you additional info about the fog if you need.
And I will check fog in DX11 plugin with RE4 today in the evening, maybe it is different 8)
Xtreme2damax,
you mean fog distance bug fixed in D3D9 ?
Or it still buggy and displaced too far ?
Check this screenshot:
http://i42.tinypic.com/2e3pr0z.jpg
-------------------------------------------
Well.... I've checked my games with Dx11 plugin:
speed is the same or a little slower in some games like Metal Gear TS.
Whole screen is pixellated, like its been rendered in native resolution without filtering...
And fog range problem is still like in other plugins...
(06-17-2010, 10:33 AM)nightrpg Wrote: [ -> ]Aside from switching between the DX9 and DX11 plugins, nothing else changes. I don't know if your plugin is just more efficiently written, DX11 is better for Dolphin than critics have forecasted it to be, or if your DX11 coding allows my video cards to take on a heavier processing burden, but whatever it is, the difference is already measurably clear.
Can't look at the video right now since my internet connection is low atm, will watch it later. Thanks for the effort so far
Anyway, THAT much of a performance boost seems a bit suspicious to me. Are you sure you're using the settings from
http://www.dolphin-emu.org/performance.html. Most notably you should check EFB copy to Texture, disable ALL XFB-related options and disable the safe texture cache (isn't supported in dx11, yet, despite being an option).
Quote:Can't look at the video right now since my internet connection is low atm, will watch it later. Thanks for the effort so far
Anyway, THAT much of a performance boost seems a bit suspicious to me. Are you sure you're using the settings from http://www.dolphin-emu.org/performance.html. Most notably you should check EFB copy to Texture, disable ALL XFB-related options and disable the safe texture cache (isn't supported in dx11, yet, despite being an option).
I just checked through my settings, and that's how I had it set up, except somehow EFB copy got switched to RAM in the DX9 plugin ><
Redoing my test, DX9 comes out ahead by about 20%