|
|
#1 | |
|
Registered User
Join Date: Jan 2005
Location: Aix en Provence (France)
Posts: 14
|
Hi,
i experienced a large FPS drop when i switched from wine-1.1.29 to wine-1.1.30, which i first reported to wine devs. The regression test i ran highlighted a change made by Henri Verbeet, which favors GL_ARB_framebuffer_object over GL_EXT_framebuffer_object when available. This had a huge impact on the frame rate for me (from 20FPS to 5FPS). After Henri's request, i reverted the code change to make wine always prefer GL_EXT_framebuffer_object, and this solved my issue. Henri Verbeet thinks : "there's probably an implementation difference between ARB_fbo and EXT_fbo in the driver". The weird thing is that i appear to be the only one person bit by the "bug". :-( I tought i may be worth it to take my chance and report it to nVidia people anyway. It happens with both binary driver 185.18.36 and 190.42. I run : - geForce 9800gt green edition with 512Mo RAM - kernel 2.6.31 - Xorg 7.1 Please let me know if can provide you with more information to investigate. I'll be glad to help. Thanks, Seb |
|
|
|
|
|
|
#2 | |
|
NVIDIA Corporation
Join Date: Mar 2005
Posts: 2,487
|
Please see the forum sticky posts about how to generate an nvidia-bug-report.log.gz. In addition, you said you tracked down the problem to a single code change in Wine. Can you please give the specific git commit ID that you're referring to, along with a detailed description of how to reproduce the problem? Thanks!
|
|
|
|
|
|
|
#3 |
|
Registered User
Join Date: Jan 2005
Location: Aix en Provence (France)
Posts: 14
|
Sorry for the lack of proper nvidia-bug-report.log file. I'll send it tonight.
In the meantime, i can give you the first "bad" commit ID : [ec97383f6fb1a14c19dfcb85dd5239a8938c9e3c] wined3d: Add support for ARB_framebuffer_object. The second one also adversely impacts the frame rate, but not as strongly : [34dd27e3a8f9affb919db6f7c912040ded8eace8] wined3d: Don't create a depthstencil renderbuffer if ARB_framebuffer_object is supported. This "bug" is reported in Wine bugzilla : http://bugs.winehq.org/show_bug.cgi?id=20307 Within is mentioned the quick fix Henri Verbeet suggested to work around the issue. Thank for your quick reply! Seb. |
|
|
|
|
|
#4 | |
|
Registered User
Join Date: Jan 2005
Location: Aix en Provence (France)
Posts: 14
|
...and I'll post the detail game setup and steps to reproduce my issue along with the bug report file.
|
|
|
|
|
|
|
#5 |
|
Registered User
Join Date: Jan 2005
Location: Aix en Provence (France)
Posts: 14
|
Here's the generated bug report, and my game options :
1680x1050@110Hz / 16:10 all options set to yes / normal BUT : -depth of field : no -soften smoke edges : no -water effect : low I launch COD4 with : WINEDEBUG=-all wine iw3mp.exe +set cg_drawfps 1 with regular wine1.1.32, i have FPS = 3-5 with the workaround suggested by Henri Verbeet, i have FPS = 15-25 Thanks for your help! |
|
|
|
|
|
#6 | |
|
Registered User
Join Date: Jan 2005
Location: Aix en Provence (France)
Posts: 14
|
@AaronP : Aaron, is the information i provided you with enough to investigate? Have you had time to take a look at it yet? Is the official 190.42 release different from the release candidate and should i try to install it?
Thanks for your help, Seb. |
|
|
|
|
|
|
#7 |
|
NVIDIA Corporation
Join Date: Mar 2005
Posts: 2,487
|
I'm sorry, I haven't yet had a chance to look at it yet. 190.42 is the same as it always was, it's just on the official driver page at nvidia.com now.
|
|
|
|
|
|
#8 |
|
Registered User
Join Date: Jan 2005
Location: Aix en Provence (France)
Posts: 14
|
@Aaron : i apologize to kind of harass you on that matter... :-(
I would really appreciate if you could give me a piece of information i could pass on to wine devs and help my "bug" make some progress... Thank you, Seb. |
|
|
|
|
|
#9 |
|
Registered User
Join Date: Jan 2005
Location: Aix en Provence (France)
Posts: 14
|
Hi,
i updated to latest 195.22 beta driver, and no longer suffer the FPS drop. Aaron, i was curious to ask : did you do something about it on the driver, or is it magically fixed "by accident" ? :-) Thanks anyway, Seb |
|
|
|
![]() |
| Thread Tools | |
|
|