http://bugs.winehq.org/show_bug.cgi?id=33478
Bug #: 33478 Summary: Borderlands 2 on FGLRX compiles a shader every frame. Product: Wine Version: unspecified Platform: x86 OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: directx-d3dx9 AssignedTo: wine-bugs@winehq.org ReportedBy: roboman2444@gmail.com CC: wine-bugs@winehq.org Classification: Unclassified
So whenever running Borderlands 2 under wine with FGLRX, I get about 2 fps. It doesn't matter the resolution, graphical settings, Borderlands 2 config file edits, or winecfg settings. Using APITrace, I have concluded that GLCompileShaderARB and GLLinkProgramARB are being called every single frame, and taking up a lot of cpu time. This may effect some other games, such as Serious Sam 2 HD, but I don't have conclusive evidence that this is the problem.
Wine version affected- 2.5.29 and below. OS- Debian Squeeze x86. CPU- amd FX8320 RAM- 24GB 1866mhz GPU- dual 6970's. Crossfire disabled for wine. Drivers- FGRLX 13.4. It has happened in previous versions as well.
http://bugs.winehq.org/show_bug.cgi?id=33478
--- Comment #1 from Alex roboman2444@gmail.com 2013-05-06 18:46:57 CDT --- It appears the HARD RESET by Flying Wild Hog also does the same thing.
http://bugs.winehq.org/show_bug.cgi?id=33478
--- Comment #2 from Alex roboman2444@gmail.com 2013-05-06 19:29:43 CDT --- I should also mention on both of those games, the console spits out stuff about gl_point being invalid. Im guessing the shader compiler is erroring on this, and trying to recompile.
http://bugs.winehq.org/show_bug.cgi?id=33478
Matteo Bruni matteo.mystral@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |matteo.mystral@gmail.com
--- Comment #3 from Matteo Bruni matteo.mystral@gmail.com 2013-05-07 09:10:45 CDT --- (In reply to comment #2)
I should also mention on both of those games, the console spits out stuff about gl_point being invalid. Im guessing the shader compiler is erroring on this, and trying to recompile.
As a start, can you attach the plain terminal log to this bug?
http://bugs.winehq.org/show_bug.cgi?id=33478
--- Comment #4 from Alex roboman2444@gmail.com 2013-05-23 21:09:44 CDT --- Sorry for the wait. The console for both Borderlands 2 and HARD RESET spit out err:d3d:state_pscale >>>>>>>>>>>>>>>>> GL_INVALID_VALUE (0x501) from glPointSize(...); @ state.c / 1591
borderlands 2 full console output (ctrl-c'd the process) http://paste.debian.net/6159/
Also, i have noticed that the HARD RESET slowdowns only occur when there is a 2d UI element being drawn. Perhaps this has to do with a render to texture shader somehow erroring on compilation?
http://bugs.winehq.org/show_bug.cgi?id=33478
Alex roboman2444@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Severity|normal |major
http://bugs.winehq.org/show_bug.cgi?id=33478
Alex roboman2444@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |roboman2444@gmail.com
http://bugs.winehq.org/show_bug.cgi?id=33478
Bruno Jesus 00cpxxx@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Severity|major |normal
--- Comment #5 from Bruno Jesus 00cpxxx@gmail.com 2013-05-25 07:17:00 CDT --- Not major, read http://bugs.winehq.org/page.cgi?id=fields.html#importance
http://bugs.winehq.org/show_bug.cgi?id=33478
jack q@cyphernaut.org changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |q@cyphernaut.org
--- Comment #6 from jack q@cyphernaut.org 2013-08-12 22:23:12 CDT --- I've experienced this issue in all Unreal Engine games that I've tested:
Borderlands 2 Remember Me Dishonored
I'd post logs but they're identical to what Alex posted earlier
I don't have extensive testing on other games (ie, non-Unreal games) but the few that I have tested don't run into any issues and run and near native Windows speed (League of Legends is the one most oft played).
OS: Arch Linux 64bit Wine: 1.7 AMD 7970M w/ Catalyst 13.8
http://bugs.winehq.org/show_bug.cgi?id=33478
Brandon Corujo haku08879@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |haku08879@gmail.com
http://bugs.winehq.org/show_bug.cgi?id=33478
--- Comment #7 from Henri Verbeet hverbeet@gmail.com 2013-08-13 02:52:35 CDT --- Does anyone have a log that actually shows the problem?
http://bugs.winehq.org/show_bug.cgi?id=33478
--- Comment #8 from jack q@cyphernaut.org 2013-08-13 05:48:11 CDT --- Here's the head of the terminal output when running Remember Me (another Unreal Engine game):
http://pastebin.mozilla.org/2846075
once it starts doing the whole:
err:d3d:wined3d_debug_callback 0x1bbac7d8: "glPointSize parameter <size> has an invalid value 'nan': must be greater than 0.000000 (GL_INVALID_VALUE)". err:d3d:state_pscale >>>>>>>>>>>>>>>>> GL_INVALID_VALUE (0x501) from glPointSize(...); @ ../../../wine/dlls/wined3d/state.c / 1579 err:d3d:wined3d_debug_callback 0x1bbac7d8: "glPointSize parameter <size> has an invalid value 'nan': must be greater than 0.000000 (GL_INVALID_VALUE)". err:d3d:state_pscale >>>>>>>>>>>>>>>>> GL_INVALID_VALUE (0x501) from glPointSize(...); @ ../../../wine/dlls/wined3d/state.c / 1579 err:d3d:wined3d_debug_callback 0x1bbac7d8: "glPointSize parameter <size> has an invalid value 'nan': must be greater than 0.000000 (GL_INVALID_VALUE)". err:d3d:state_pscale >>>>>>>>>>>>>>>>> GL_INVALID_VALUE (0x501) from glPointSize(...); @ ../../../wine/dlls/wined3d/state.c / 1579 err:d3d:wined3d_debug_callback 0x1bbac7d8: "glPointSize parameter <size> has an invalid value 'nan': must be greater than 0.000000 (GL_INVALID_VALUE)". err:d3d:state_pscale >>>>>>>>>>>>>>>>> GL_INVALID_VALUE (0x501) from glPointSize(...); @ ../../../wine/dlls/wined3d/state.c / 1579
it doesn't stop until I quit the game. Strangely enough, it ran better when I ran it to pull this output, but still unplayable (< 10fps)
http://bugs.winehq.org/show_bug.cgi?id=33478
--- Comment #9 from jack q@cyphernaut.org 2013-08-13 05:56:29 CDT --- And for the sake of completeness, here's the output of running Borderlands 2 (it's largely the same, same repeating error message):
http://pastebin.mozilla.org/2846106
And here's what I can pull out of stracing the Borderlands2 pid:
http://pastebin.mozilla.org/2846107
again, tons of repeating stuff out of this
http://bugs.winehq.org/show_bug.cgi?id=33478
--- Comment #10 from jack q@cyphernaut.org 2013-08-13 05:59:37 CDT --- http://bugs.winehq.org/show_bug.cgi?id=23048
http://bugs.winehq.org/show_bug.cgi?id=33478
--- Comment #11 from Henri Verbeet hverbeet@gmail.com 2013-08-13 06:25:50 CDT --- (In reply to comment #8)
err:d3d:wined3d_debug_callback 0x1bbac7d8: "glPointSize parameter <size> has an invalid value 'nan': must be greater than 0.000000 (GL_INVALID_VALUE)".
Those are essentially bug 22922 / bug 23048. Making them go away would require some tests to see what is supposed to happen when the applications sets NAN as point size.
Those messages have nothing to do with shader compilation though, which is what this bug claims to be about.
http://bugs.winehq.org/show_bug.cgi?id=33478
--- Comment #11 from Henri Verbeet hverbeet@gmail.com 2013-08-13 06:25:50 CDT --- (In reply to comment #8)
err:d3d:wined3d_debug_callback 0x1bbac7d8: "glPointSize parameter <size> has an invalid value 'nan': must be greater than 0.000000 (GL_INVALID_VALUE)".
Those are essentially bug 22922 / bug 23048. Making them go away would require some tests to see what is supposed to happen when the applications sets NAN as point size.
Those messages have nothing to do with shader compilation though, which is what this bug claims to be about.
--- Comment #12 from Alex roboman2444@gmail.com 2013-08-13 12:39:02 CDT --- (In reply to comment #11)
(In reply to comment #8)
err:d3d:wined3d_debug_callback 0x1bbac7d8: "glPointSize parameter <size> has an invalid value 'nan': must be greater than 0.000000 (GL_INVALID_VALUE)".
Those are essentially bug 22922 / bug 23048. Making them go away would require some tests to see what is supposed to happen when the applications sets NAN as point size.
Those messages have nothing to do with shader compilation though, which is what this bug claims to be about.
Sure, that is a bug, but it isnt the bug which slows things down to 2 fps. If you actually look at traces, you would notice that the two calls that take the most amount of cpu time in between each flip are directly related to compiling and linking a shader program.
Please do some research before commenting. Just reading logs is not enough.
http://bugs.winehq.org/show_bug.cgi?id=33478
--- Comment #13 from Henri Verbeet hverbeet@gmail.com 2013-08-13 12:50:58 CDT --- (In reply to comment #12)
(In reply to comment #11)
Those messages have nothing to do with shader compilation though, which is what this bug claims to be about.
Sure, that is a bug, but it isnt the bug which slows things down to 2 fps. If you actually look at traces,
I'd consider looking at those traces you mention if any of them where actually attached to this bug.
Please do some research before commenting. Just reading logs is not enough.
Yeah, good luck with that.
http://bugs.winehq.org/show_bug.cgi?id=33478
--- Comment #14 from Alex roboman2444@gmail.com 2013-08-13 13:00:55 CDT --- (In reply to comment #13)
(In reply to comment #12)
(In reply to comment #11)
I'd consider looking at those traces you mention if any of them where actually attached to this bug.
I don't have any AMD cards with me anymore. Someone else will have to get traces.
http://bugs.winehq.org/show_bug.cgi?id=33478
--- Comment #15 from jack q@cyphernaut.org 2013-09-04 13:58:51 CDT --- This set of patches seems to have fixed my issues:
http://www.winehq.org/pipermail/wine-devel/2013-September/101106.html
This is 100% unscientific, but after applying those patches and disabling StrictDrawOrdering Borderlands2 runs flawlessly with fglrx.
https://bugs.winehq.org/show_bug.cgi?id=33478
--- Comment #16 from Austin English austinenglish@gmail.com --- This is your friendly reminder that there has been no bug activity for over a year. Is this still an issue in current (1.7.37 or newer) wine? If so, please attach the terminal output in 1.7.37 (see http://wiki.winehq.org/FAQ#get_log).
https://bugs.winehq.org/show_bug.cgi?id=33478
Matteo Bruni matteo.mystral@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Component|directx-d3dx9 |directx-d3d
https://bugs.winehq.org/show_bug.cgi?id=33478
joaopa jeremielapuree@yahoo.fr changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |jeremielapuree@yahoo.fr
--- Comment #17 from joaopa jeremielapuree@yahoo.fr --- No answer since for more 5 years, no download. This bug can be closed as ABANDONED.
https://bugs.winehq.org/show_bug.cgi?id=33478
Matteo Bruni matteo.mystral@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Resolution|--- |ABANDONED Status|UNCONFIRMED |RESOLVED
--- Comment #18 from Matteo Bruni matteo.mystral@gmail.com --- fglrx is not a thing anymore and it seems unlikely that anything of significance will ever come out of this bug.
https://bugs.winehq.org/show_bug.cgi?id=33478
André H. nerv@dawncrow.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED CC| |nerv@dawncrow.de
--- Comment #19 from André H. nerv@dawncrow.de --- closing abandoned