http://bugs.winehq.org/show_bug.cgi?id=36489
Bug ID: 36489 Summary: Fate:Stay/Night performance regression Product: Wine Version: unspecified Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown Assignee: wine-bugs@winehq.org Reporter: elfmax@tut.by CC: stefan@codeweavers.com Regression SHA1: 19f45af19cd0c1162f15d6779506431d1cd99d8c
Created attachment 48561 --> http://bugs.winehq.org/attachment.cgi?id=48561 Regression testing result
Drastically decreased performance in Fate:Stay/Night Visual Novel when upgrading from 1.7.18 to 1.7.19.
The game used to run smoothly on 1.7.18 after tweaking around userconf.exe, there was (not english) menu choice between BitBlt/dibsect,DrawDib/drawdib and drawdibjdi and StretchDiBits/sdb, the first being the fastest.
This game, being a 2-D visual novel, operates mainly on 2-D images, moving/resizing/applying effects to them.
http://bugs.winehq.org/show_bug.cgi?id=36489
Dennis Borisevich elfmax@tut.by changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |elfmax@tut.by
http://bugs.winehq.org/show_bug.cgi?id=36489
--- Comment #1 from Stefan Dösinger stefan@codeweavers.com --- Was this game slow before patch adf0ab2ad889c7c6bcda484edfde9a431f829b08 ?
http://bugs.winehq.org/show_bug.cgi?id=36489
Henri Verbeet hverbeet@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |regression Component|-unknown |directx-d3d Version|unspecified |1.7.19
http://bugs.winehq.org/show_bug.cgi?id=36489
--- Comment #2 from Dennis Borisevich elfmax@tut.by --- (In reply to Stefan Dösinger from comment #1)
Was this game slow before patch adf0ab2ad889c7c6bcda484edfde9a431f829b08 ?
I checked out f6403b4ac0ad822b3c1b4e9134b4d5cd2dfce6c6, which was right before the patch you mentioned. The problem is present, game is slow.
http://bugs.winehq.org/show_bug.cgi?id=36489
--- Comment #3 from Stefan Dösinger stefan@codeweavers.com --- 19f45af19cd0c1162f15d6779506431d1cd99d8c reverts an optimization that was part of adf0ab2ad889c7c6bcda484edfde9a431f829b08 because that optimization caused correctness problems (see bug 36227). This correctness problem was triggered by a broken game.
You should be able to work around the problem by setting HKCU/Software/Wine/Direct3D/DirectDrawRenderer = gdi. In this case we use DIB sections as the main store for DirectDraw surfaces.
I'll have a look at Fate:Stay/Night Visual Novel and see what it does. Since it offers the user a choice of how it uses the Windows API I suspect it is doing something questionable as well. Maybe I can find a way to make both games happy.
http://bugs.winehq.org/show_bug.cgi?id=36489
Dennis Borisevich elfmax@tut.by changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |RESOLVED Resolution|--- |WONTFIX
--- Comment #4 from Dennis Borisevich elfmax@tut.by --- (In reply to Stefan Dösinger from comment #3)
19f45af19cd0c1162f15d6779506431d1cd99d8c reverts an optimization that was part of adf0ab2ad889c7c6bcda484edfde9a431f829b08 because that optimization caused correctness problems (see bug 36227). This correctness problem was triggered by a broken game.
You should be able to work around the problem by setting HKCU/Software/Wine/Direct3D/DirectDrawRenderer = gdi. In this case we use DIB sections as the main store for DirectDraw surfaces.
I'll have a look at Fate:Stay/Night Visual Novel and see what it does. Since it offers the user a choice of how it uses the Windows API I suspect it is doing something questionable as well. Maybe I can find a way to make both games happy.
It works better with this registry key set to "gdi". Actually it runs better than before ( no black artifacts on startup). Thanks for advice.
http://bugs.winehq.org/show_bug.cgi?id=36489
--- Comment #5 from Henri Verbeet hverbeet@gmail.com --- Note that the "DirectDrawRenderer" registry key will eventually go away.
http://bugs.winehq.org/show_bug.cgi?id=36489
Stefan Dösinger stefan@codeweavers.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|WONTFIX |--- Ever confirmed|0 |1
--- Comment #6 from Stefan Dösinger stefan@codeweavers.com --- I wouldn't mark this bug WONTFIX just yet. While it will be tricky to fix this bug and bug 36227 at the same time, there are still some options I'd like to look into. Maybe we can use the hSection/dwOffset parameters of CreateDIBSection to create the DIB section around the already allocated system memory. Option two is to ignore SFLAG_PIN_SYSMEM if the application has never locked the surface before it called GetDC.
http://bugs.winehq.org/show_bug.cgi?id=36489
--- Comment #7 from Stefan Dösinger stefan@codeweavers.com --- Is there a demo of this game? I found an apparent demo for the GameBoy Advance, and there are articles claiming that there is a PS Vita demo, but I found no demo for the Windows game.
If there is no demo, can you please attach a +ddraw log?
http://bugs.winehq.org/show_bug.cgi?id=36489
--- Comment #8 from Dennis Borisevich elfmax@tut.by --- Created attachment 48670 --> http://bugs.winehq.org/attachment.cgi?id=48670 +ddraw log with GDI enabled
Attached gzip'd log with GDI enabled as the main renderer.
http://bugs.winehq.org/show_bug.cgi?id=36489
--- Comment #9 from Dennis Borisevich elfmax@tut.by --- Created attachment 48671 --> http://bugs.winehq.org/attachment.cgi?id=48671 +ddraw log with OpenGL enabled
Attached gzip'd log with OpenGL DirectDraw rendering.
On both startups I did simple things: - waited for main menu to open - loaded a game - picked through some lines of text in VN - closed game - CTRL-C stopped the game from console ( always like that, it does not close wine window, although seems to exit )
https://bugs.winehq.org/show_bug.cgi?id=36489
--- Comment #10 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 (1.7.51 or newer) wine?
https://bugs.winehq.org/show_bug.cgi?id=36489
--- Comment #11 from Dennis Borisevich elfmax@tut.by --- Hello, on new hardware and fresh wineprefix (wine version 1.7.51) the bug is not present, nevermind the intro credits (a couple of white surfaces are actually black, which is noticeable).
https://bugs.winehq.org/show_bug.cgi?id=36489
--- Comment #12 from Austin English austinenglish@gmail.com --- (In reply to Dennis Borisevich from comment #11)
Hello, on new hardware and fresh wineprefix (wine version 1.7.51) the bug is not present
Thanks, marking abandoned since you're the only reporter here.
nevermind the intro credits (a couple of white surfaces are actually black, which is noticeable).
Please file a separate bug report for that.
https://bugs.winehq.org/show_bug.cgi?id=36489
Bruno Jesus 00cpxxx@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution|--- |ABANDONED
--- Comment #13 from Bruno Jesus 00cpxxx@gmail.com --- (In reply to Austin English from comment #12)
(In reply to Dennis Borisevich from comment #11)
Hello, on new hardware and fresh wineprefix (wine version 1.7.51) the bug is not present
Thanks, marking abandoned since you're the only reporter here.
https://bugs.winehq.org/show_bug.cgi?id=36489
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #14 from Austin English austinenglish@gmail.com --- Closing.
-
wine-bugs＠winehq.org