[Bug 34542] New: Regression - Kega Fusion 3.64 - Slow directx performance and fullscreen not working anymore.
http://bugs.winehq.org/show_bug.cgi?id=34542
Bug #: 34542 Summary: Regression - Kega Fusion 3.64 - Slow directx performance and fullscreen not working anymore. Product: Wine Version: 1.7.2 Platform: x86 OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown AssignedTo: wine-bugs@winehq.org ReportedBy: joaogojunior@hotmail.com Classification: Unclassified
I can't get fullscreen on wine-1.7.2, but it has been working perfectly on wine-1.6, also noticed this on other directx applications (snesx9 v1.51, phantasy star online v2). On windowed mode kega fusion loads and runs but with an degraded performance of 4 to 5 fps, compared to 60 fps on wine-1.6 on the same machine... My system is an Slackware 13.37 32bits, latest nvidia legacy drivers (96.43.23) on geforce mx 4000 64MB, AthlonXP and 1GB DDR1 RAM.
http://bugs.winehq.org/show_bug.cgi?id=34542
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |performance, regression Summary|Regression - Kega Fusion |Kega Fusion 3.64 - Slow |3.64 - Slow directx |directx performance |performance and fullscreen | |not working anymore. |
--- Comment #1 from Austin English austinenglish@gmail.com 2013-09-18 14:37:46 CDT --- (In reply to comment #0)
I can't get fullscreen on wine-1.7.2, but it has been working perfectly on wine-1.6, also noticed this on other directx applications (snesx9 v1.51, phantasy star online v2). On windowed mode kega fusion loads and runs but with an degraded performance of 4 to 5 fps, compared to 60 fps on wine-1.6 on the same machine...
Please run a regression test: http://wiki.winehq.org/RegressionTesting
and file a separate bug for the fullscreen problem (you'll want to run a separate regression test for that problem).
http://bugs.winehq.org/show_bug.cgi?id=34542
--- Comment #2 from João Guilherme joaogojunior@hotmail.com 2013-09-22 07:42:21 CDT --- Better testing showed that everything was working allright on wine-1.7.1... Normal windowed performance, I'll fill in another bug on fullscreen problem but on wine-1.7.1 does not display that problem... I'll start regression testing from 1.7.1 to 1.7.2... Thanks alot!
http://bugs.winehq.org/show_bug.cgi?id=34542
--- Comment #3 from João Guilherme joaogojunior@hotmail.com 2013-09-23 04:30:52 CDT --- Did the bissection, Here is the result:
6f321ff2f97a4eb1e613022342399c3ffc41bc89 is the first bad commit commit 6f321ff2f97a4eb1e613022342399c3ffc41bc89 Author: Henri Verbeet hverbeet@codeweavers.com Date: Thu Sep 5 09:54:39 2013 +0200
ddraw: Don't imply WINED3DUSAGE_RENDERTARGET on primary surfaces.
If the application actually wants to do 3D rendering on the primary surface, it should set DDSCAPS_3DDEVICE as well.
:040000 040000 7489dc38430c1447071ecac4bbe3d21cfdb3b7f8 4fe57e7d8e75952d72c0916dca3107b22de5f8ee M dlls
If this patch is reverted, kega fusion runs smooth on windowed mode... (fullscreen works fine with or without this patch).
http://bugs.winehq.org/show_bug.cgi?id=34542
Henri Verbeet hverbeet@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Regression SHA1| |6f321ff2f97a4eb1e6130223423 | |99c3ffc41bc89
--- Comment #4 from Henri Verbeet hverbeet@gmail.com 2013-09-23 04:35:26 CDT --- Could you attach a "+seh,+tid,+ddraw,+d3d,+d3d_surface" log from before the regression, and one with current git?
http://bugs.winehq.org/show_bug.cgi?id=34542
--- Comment #5 from João Guilherme joaogojunior@hotmail.com 2013-09-23 05:13:43 CDT --- Created attachment 46035 --> http://bugs.winehq.org/attachment.cgi?id=46035 Kega fusion on current git, after windowed performance regression
http://bugs.winehq.org/show_bug.cgi?id=34542
--- Comment #6 from João Guilherme joaogojunior@hotmail.com 2013-09-23 05:19:02 CDT --- Created attachment 46036 --> http://bugs.winehq.org/attachment.cgi?id=46036 Kega fusion on wine-1.7.1 before windowed performance regression
http://bugs.winehq.org/show_bug.cgi?id=34542
João Guilherme joaogojunior@hotmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |hverbeet@gmail.com
http://bugs.winehq.org/show_bug.cgi?id=34542
--- Comment #7 from João Guilherme joaogojunior@hotmail.com 2013-09-23 05:46:52 CDT --- Thanks Henri Verbeet, for your time for looking into this, if you need something else just ask...
http://bugs.winehq.org/show_bug.cgi?id=34542
--- Comment #8 from Henri Verbeet hverbeet@gmail.com 2013-09-23 05:55:58 CDT --- Created attachment 46037 --> http://bugs.winehq.org/attachment.cgi?id=46037 patch
Does the attached patch help? If not, could you please attach a new log with the patch applied on top of current git?
http://bugs.winehq.org/show_bug.cgi?id=34542
--- Comment #9 from João Guilherme joaogojunior@hotmail.com 2013-09-23 08:13:42 CDT --- The patch did help! the performance on window mode is great again! but my knowledge still very limited I see that you have removed a block of code, I am wondering if this will make diference for other programs? I did generate the log still interested?
http://bugs.winehq.org/show_bug.cgi?id=34542
--- Comment #10 from Henri Verbeet hverbeet@gmail.com 2013-09-23 10:14:54 CDT --- (In reply to comment #9)
The patch did help! the performance on window mode is great again! but my knowledge still very limited I see that you have removed a block of code, I am wondering if this will make diference for other programs? I did generate the log still interested?
surface_blt_special() contains some optimized paths for certain types of blits. The if-statement that is removed by the patch restricts it to render targets, probably to avoid a bunch of checks if we already know surface_blt_special() can't help. Unfortunately the if-statement is incorrect, because the restriction should really be about onscreen surfaces. It's also redundant because surface_blt_special() already contains a correct check at http://source.winehq.org/git/wine.git/blob/24fbe9f2ecfdac12dcf7af49f03d7624d....
The reason you're hitting this code at all is because your GL driver doesn't support FBO blits, and the hardware doesn't support fragment shaders, which would have been the preferred ways of doing blits.
http://bugs.winehq.org/show_bug.cgi?id=34542
Henri Verbeet hverbeet@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Fixed by SHA1| |ea4c3d5d5d3df37391a349542ce | |82b5a4f56d470 Status|UNCONFIRMED |RESOLVED Resolution| |FIXED
--- Comment #11 from Henri Verbeet hverbeet@gmail.com 2013-09-25 02:59:55 CDT --- Should be fixed by commit ea4c3d5d5d3df37391a349542ce82b5a4f56d470.
http://bugs.winehq.org/show_bug.cgi?id=34542
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #12 from Alexandre Julliard julliard@winehq.org 2013-09-27 13:41:16 CDT --- Closing bugs fixed in 1.7.3.
https://bugs.winehq.org/show_bug.cgi?id=34542
Anastasius Focht focht@gmx.net changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |focht@gmx.net Component|-unknown |directx-d3d
-
wine-bugs＠winehq.org