http://bugs.winehq.org/show_bug.cgi?id=12453
Summary: Counter Strike Source bad performance in Video Stress Test Product: Wine Version: CVS/GIT Platform: PC URL: http://appdb.winehq.org/objectManager.php?sClass=version &iId=3731 OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: directx-d3d AssignedTo: wine-bugs@winehq.org ReportedBy: vitaliy@kievinfo.com
During almost entire video stress test FPS stay really low (several times less the on windows). This happens while console is being filled with these messages:
fixme:d3d:IWineD3DCubeTextureImpl_PreLoad Cubetexture (0x1dfbcda8) has been reloaded at least 20 times due to WINED3DSAMP_SRGBTEXTURE changes on it's sampler err:d3d_surface:fb_copy_to_texture_direct Texture filtering not supported in direct blit
http://bugs.winehq.org/show_bug.cgi?id=12453
Paul BerlinerPaul@web.de changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |BerlinerPaul@web.de
--- Comment #1 from Paul BerlinerPaul@web.de 2008-09-19 19:25:20 --- confirming. the test lags or freeze the whole computer.
http://bugs.winehq.org/show_bug.cgi?id=12453
Vitaliy Margolen vitaliy@kievinfo.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |thebozo1@hotmail.com
--- Comment #2 from Vitaliy Margolen vitaliy@kievinfo.com 2008-12-29 09:35:54 --- *** Bug 14451 has been marked as a duplicate of this bug. ***
http://bugs.winehq.org/show_bug.cgi?id=12453
Vitaliy Margolen vitaliy@kievinfo.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |ambro@b4ever.net
--- Comment #3 from Vitaliy Margolen vitaliy@kievinfo.com 2008-12-29 09:36:35 --- *** Bug 12993 has been marked as a duplicate of this bug. ***
http://bugs.winehq.org/show_bug.cgi?id=12453
Vitaliy Margolen vitaliy@kievinfo.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Severity|normal |major Summary|Counter Strike Source bad |Bad performance Source based |performance in Video Stress |games |Test | Version|CVS/GIT |0.9.60
--- Comment #4 from Vitaliy Margolen vitaliy@kievinfo.com 2008-12-29 09:41:25 --- Changing summary to more generic.
http://bugs.winehq.org/show_bug.cgi?id=12453
Anton Romanov theli@ukr.net changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |theli@ukr.net
http://bugs.winehq.org/show_bug.cgi?id=12453
Conrad Bebbington conrad@bebbington.org changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |conrad@bebbington.org
http://bugs.winehq.org/show_bug.cgi?id=12453
Night Nord NightNord@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |NightNord@gmail.com
http://bugs.winehq.org/show_bug.cgi?id=12453
Artem S. Tashkinov t.artem@mailcity.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |t.artem@mailcity.com
http://bugs.winehq.org/show_bug.cgi?id=12453
Gary Pendergast winehq@pento.net changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |winehq@pento.net
http://bugs.winehq.org/show_bug.cgi?id=12453
Denis Misiurca infoman1985@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |infoman1985@gmail.com
http://bugs.winehq.org/show_bug.cgi?id=12453
Slash Bunny demodevil@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |demodevil@gmail.com
http://bugs.winehq.org/show_bug.cgi?id=12453
Jaime Rave jaimerave@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |jaimerave@gmail.com
http://bugs.winehq.org/show_bug.cgi?id=12453
dlzerocool dl.zerocool@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |dl.zerocool@gmail.com
--- Comment #5 from dlzerocool dl.zerocool@gmail.com 2009-01-01 12:04:55 --- Hello,
After a lot of tests I can conclude that the lightning system is the problem of low framerate.
Actually when you set your game to a resolution then set it again to a newer resolution for exemple from 1440x900 to 9XXxXXX the lights in game change dramatically.
And strangely you get a real boost in FPS. Everything looks more darken but it do something that makes wine manage in another way the lightnings.
Hope this can help in a way to found the bug.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #6 from H. Verbeet hverbeet@gmail.com 2009-01-01 12:15:24 --- No, for the most part the problem is sRGB textures.
http://bugs.winehq.org/show_bug.cgi?id=12453
chris ahrendt celticht32@aol.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |celticht32@aol.com
http://bugs.winehq.org/show_bug.cgi?id=12453
Jim Newland jnewl@socal.rr.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |jnewl@socal.rr.com
--- Comment #7 from Jim Newland jnewl@socal.rr.com 2009-01-02 13:47:33 --- I have nothing to add except to report that the game works fine under Cedega (7.0.0 engine) with no slowdown. I will list the Cedega graphics options and their settings here, as they may give a clue as to what needs to be done to get the game running properly in Wine (I myself have no idea; just trying to help):
ENGINE: Cedega engine version: 7.0.0 Windows verion: Windows XP Restrict to single core: disabled Thread delay (ms): 10 Scheduler & Decrease Server Priority: both disabled Accelerated Interprocess Communication: enabled
WINDOW SYSTEM: Run in Window: no Resolution Control: XVidMode Managed: enabled Activate Cedega HUD: Disabled Lock Mouse to Game: enabled Enable FreeType and XRender: enabled
GRAPHICS: Video RAM: 512 (my card is an Nvidia GeForce 8800GT) AGP Vertex Data (MB): Autodetect GL Shading Language: Mixed Pixel Shaders: 3.0, enabled Vertex Shaders: 3.0, enabled Fixed Program: Auto Fragment Offset: Auto Non Power of Two Textures: Auto Deprecated Features: disabled Fixed GL Extension Buffer: disabled Manual GL Extensions: none Clip Space Fix: enabled Anisotropic Filtering: disabled Framebuffer Objects & Float Textures: both enabled Occlusion Queries: enabled NV_VAR Extension: disabled (note says it takes priority over ARB_VBO when both are enabled; this *must* be disabled for Nvidia users in order to play game.) ARB_VBO Extension: enabled Dynamic VBO: disabled Index VBO: disabled
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #8 from H. Verbeet hverbeet@gmail.com 2009-01-02 14:35:29 --- Created an attachment (id=18409) --> (http://bugs.winehq.org/attachment.cgi?id=18409) Small sRGB hack
The attached patch should give a significant performance improvement to most source engine based games. It's also obviously incorrect, of course.
http://bugs.winehq.org/show_bug.cgi?id=12453
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |patch
http://bugs.winehq.org/show_bug.cgi?id=12453
Vitaliy Margolen vitaliy@kievinfo.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords|patch | Summary|Bad performance Source based|Bad performance in Source |games |based games
--- Comment #9 from Vitaliy Margolen vitaliy@kievinfo.com 2009-01-02 15:37:12 --- (In reply to comment #7)
I have nothing to add except to report that the game works fine under Cedega (7.0.0 engine)
Please don't add such "information" here. Cedega has absolutely nothing to do with Wine when it comes to d3d.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #10 from dlzerocool dl.zerocool@gmail.com 2009-01-04 09:05:41 --- (In reply to comment #8)
Created an attachment (id=18409)
--> (http://bugs.winehq.org/attachment.cgi?id=18409) [details]
Small sRGB hack
The attached patch should give a significant performance improvement to most source engine based games. It's also obviously incorrect, of course.
I tried this patch, and what I get is the game being unable to run.
I didn't test to compile whitout (In reply to comment #8)
Created an attachment (id=18409)
--> (http://bugs.winehq.org/attachment.cgi?id=18409) [details]
Small sRGB hack
The attached patch should give a significant performance improvement to most source engine based games. It's also obviously incorrect, of course.
Good thing, except that at moment I can't test it. Actual GIT doesn't run l4d, steam run nicely but I can't start left 4 dead. This is not due to your patch, I've compile it with and without your patch, it's a current git regression.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #11 from dlzerocool dl.zerocool@gmail.com 2009-01-04 09:06:53 --- This part of the comment is wrong sorry :
" (In reply to comment #10)
(In reply to comment #8)
Created an attachment (id=18409)
--> (http://bugs.winehq.org/attachment.cgi?id=18409) [details] [details]
Small sRGB hack
The attached patch should give a significant performance improvement to most source engine based games. It's also obviously incorrect, of course.
I tried this patch, and what I get is the game being unable to run.
I didn't test to compile whitout (In reply to comment #8)
"
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #12 from Vitaliy Margolen vitaliy@kievinfo.com 2009-01-04 12:37:52 --- Hack works for me. In CSS Stress test (1024x768, DX90, max settings on 8800GT): without patch: 82 FPS with patch: 172 FPS
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #13 from Denis Misiurca infoman1985@gmail.com 2009-01-04 13:42:27 --- The patch works for Left4Dead, but TF2 looks sometimes ugly with it. I'll try to make it optional per-application today or tomorrow.
http://bugs.winehq.org/show_bug.cgi?id=12453
Paul "TBBle" Hampson Paul.Hampson@Pobox.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |Paul.Hampson@Pobox.com
--- Comment #14 from Paul "TBBle" Hampson Paul.Hampson@Pobox.com 2009-01-06 00:07:59 --- (In reply to comment #8)
Created an attachment (id=18409)
--> (http://bugs.winehq.org/attachment.cgi?id=18409) [details]
Small sRGB hack
The attached patch should give a significant performance improvement to most source engine based games. It's also obviously incorrect, of course.
If I'm following that patch correctly, then it simply disables support for the SRGB_TEXTURE sampler state, which seems to overly brighten everything when applied to Warhammer Online (which seems to suffer the same bug except in texture, not cubetexture.). I guess that makes sense, since it's disabling the sRGB->linear conversion in the texture sampler, since linear values are lower than the sRGB value for the same colour. (Assuming I've got sRGB correct...)
If I'm understanding the code and specs correctly, then the problem is that D3D lets the sampler have sRGB->linear conversion turned on and off whenever it likes, while OpenGL requires a reload of the relevant texture with a different format.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #15 from H. Verbeet hverbeet@gmail.com 2009-01-06 06:45:02 --- (In reply to comment #14)
If I'm following that patch correctly, then it simply disables support for the SRGB_TEXTURE sampler state, which seems to overly brighten everything when applied to Warhammer Online (which seems to suffer the same bug except in texture, not cubetexture.). I guess that makes sense, since it's disabling the sRGB->linear conversion in the texture sampler, since linear values are lower than the sRGB value for the same colour. (Assuming I've got sRGB correct...)
If I'm understanding the code and specs correctly, then the problem is that D3D lets the sampler have sRGB->linear conversion turned on and off whenever it likes, while OpenGL requires a reload of the relevant texture with a different format.
Yeah, that's pretty much the problem. We could speed it up a bit by duplicating the textures and having a regular and sRGB version, but that does mean using twice the memory for some textures. Unfortunately it's not like the driver has lots of address space to spare either.
http://bugs.winehq.org/show_bug.cgi?id=12453
Chris Boyle chris@boyle.name changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |chris@boyle.name
--- Comment #16 from Chris Boyle chris@boyle.name 2009-01-07 08:36:24 --- Left 4 Dead without patch: ~6 fps. With patch: ~19fps. Thanks. :-)
http://bugs.winehq.org/show_bug.cgi?id=12453
Gregor Münch greg87@online.de changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |greg87@online.de
--- Comment #17 from Gregor Münch greg87@online.de 2009-01-19 08:11:02 --- For me dx9.0 mode is unplayable slowly and with the hack it looks ugly. CS:S Stresstest: Wine 1024x768 ~40fps (Note: The stresstest doesnt run with 1440x900 in wine, it just freezes) Windows 1440x900 ~200fps+
I use now dx8.1 mode. And if fbo is selected in registry I get a nice speed boost. The stresstest went up from ~80fps to ~120fps.
http://bugs.winehq.org/show_bug.cgi?id=12453
Tobias Jakobi liquid.acid@gmx.net changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |liquid.acid@gmx.net
http://bugs.winehq.org/show_bug.cgi?id=12453
Jan Buecken jb.faq@gmx.de changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |jb.faq@gmx.de
http://bugs.winehq.org/show_bug.cgi?id=12453
bart nagel bart+bugs.winehq.org@tremby.net changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |bart+bugs.winehq.org@tremby. | |net
http://bugs.winehq.org/show_bug.cgi?id=12453
DL taedium_vitae@eml.cc changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |taedium_vitae@eml.cc
--- Comment #18 from DL taedium_vitae@eml.cc 2009-02-18 03:36:25 --- Seems to be fixed in current git, at least with HL2, which now runs at 60+ fps instead of 20+ fps for wine-1.1.15.Left4Dead doesn't seem to have any noticeable performance increase in between current git and wine-1.1.15; both get between 20-38fps or so with low settings.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #19 from Tobias Jakobi liquid.acid@gmx.net 2009-02-18 05:47:54 --- (In reply to comment #18)
Seems to be fixed in current git, at least with HL2, which now runs at 60+ fps instead of 20+ fps for wine-1.1.15.Left4Dead doesn't seem to have any noticeable performance increase in between current git and wine-1.1.15; both get between 20-38fps or so with low settings.
That's because of stefand's recent commit: http://source.winehq.org/git/wine.git/?a=commit;h=c585b4de9949ec3cb583de4a3a...
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #20 from Paul "TBBle" Hampson Paul.Hampson@Pobox.com 2009-02-18 08:09:18 --- A build of GIT head (bd69c2e) produces a continuous output of the following (with different addresses) from Warhammer Online
fixme:d3d_surface:IWineD3DSurfaceImpl_LoadLocation 0x6831818: Downloading rgb texture to reload it as srgb
I also see: fixme:d3d_surface:surface_load_ds_location (0x162860) Not supported with fixed up depth stencil occasionally, followed by a few GL_OUT_OF_MEMORY errors, every time a certain address reports the first error in this comment.
Reducing available texture memory to 128MB doesn't fix the GL_OUT_OF_MEMORY errors, but switching OffScreenRenderingMode from fbo to backbuffer did make the latter error disappear. In all combinations of these two changes the first error in this comment occurred a lot.
I can't speak for its performance compared to before the patch either.
If I'm skimming the code correctly, the issue causing this is that the texture has been uploaded as an RGB texture and removed from system memory (or was never there... I don't know if that can happen under Direct3D) and now is being asked for an an SRGB texture, so has to pull the RGB texture down into system memory and upload it again as SRGB.
I'm not sure that this is not due to Warhammer Online shuffling textures around due to low video memory availability (256MB card) since the only other machine I have access to for Warhammer Online under Wine dereferences a NULL inside WAR.exe (chasing that up with EA support separately) before getting as far as rendering. That would mean that the patch is fine as it is, but the minimum video card memory for games that use textures in both RGB and SRGB from the same sampler turns out to be much higher under Wine than Windows. That's prolly worth noting somewhere.
It might be useful to also have an updated patch that disables changing the SRGB-ness of a sampler after the texture is loaded. (Which is what the patch on the Warhammer AppDB page effectively does) Bizarrely, as I recall that looked correct under Warhammer Online, suggesting that all the textures were being loaded correctly the first time, and then having their state switched repeatedly but were only ever rendered in the right state.
I wonder if it's possible to delay the re-upload of a texture until closer to rendering, to catch what I'm guessing is a sampler change that happens twice per frame... (I really should test this)
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #21 from DL taedium_vitae@eml.cc 2009-02-23 02:42:10 ---
fixme:d3d_surface:surface_load_ds_location (0x162860) Not supported with fixed
up depth stencil
Yeah, I get this message with both Portal and Left4Dead; both have similar performance to wine-1.1.15.Half-Life 2 is the only source game so far that does not have this message, and so has much improved performance.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #22 from DL taedium_vitae@eml.cc 2009-02-23 02:52:42 --- (In reply to comment #20)
It might be useful to also have an updated patch that disables changing the SRGB-ness of a sampler after the texture is loaded. (Which is what the patch on the Warhammer AppDB page effectively does) Bizarrely, as I recall that looked correct under Warhammer Online, suggesting that all the textures were being loaded correctly the first time, and then having their state switched repeatedly but were only ever rendered in the right state.
Could you post a link to the patch on the Warhammer AppDB page? I see the vertex blending hack, but this is not related to this issue, correct?
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #23 from Tobias Jakobi liquid.acid@gmx.net 2009-02-23 05:17:04 --- (In reply to comment #21)
fixme:d3d_surface:surface_load_ds_location (0x162860) Not supported with fixed
up depth stencil
Yeah, I get this message with both Portal and Left4Dead; both have similar performance to wine-1.1.15.Half-Life 2 is the only source game so far that does not have this message, and so has much improved performance.
Confirming with L4D on a FX5900. I have turned all graphics settings down to low but this FIXME remains (using git master here).
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #24 from Paul "TBBle" Hampson Paul.Hampson@Pobox.com 2009-02-24 06:27:46 --- (In reply to comment #22)
(In reply to comment #20)
It might be useful to also have an updated patch that disables changing the SRGB-ness of a sampler after the texture is loaded. (Which is what the patch on the Warhammer AppDB page effectively does)
Could you post a link to the patch on the Warhammer AppDB page? I see the vertex blending hack, but this is not related to this issue, correct?
I think it's the patch in the howto. There's two parts, one is the vertex blending hack, the other removes the SRGB reloading. However, that code has been completely removed by the patch addressing this bug (mentioned in comment 18), so needs to be redesigned to fit into the changed code.
http://bugs.winehq.org/show_bug.cgi?id=12453
James Henderson james.m.henderson@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |james.m.henderson@gmail.com
--- Comment #25 from James Henderson james.m.henderson@gmail.com 2009-02-24 18:43:06 --- I compiled using recent git (wine-1.1.15-221-g79cc416) on my x86_64 debian lenny based system using nvidia 8600gt video card.
The IWineD3DCubeTextureImpl message does not seem to be occuring anymore.
However, I am getting a lot of the following error: fixme:d3d_surface:IWineD3DVolumeImpl_LockBox (0x27c63930) : pBox=0x33de64 stub
As well as very slow framerate (unplayable).
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #26 from DL taedium_vitae@eml.cc 2009-02-24 22:59:35 --- Okay, I tried the SRGB hack from the Warhammer Online appdb page and applied it against wine 1.1.15.It gives quite a nice performance boost compared to wine-1.1.15 for left4dead,portal and hl2 although the fps increase is not as great as the increase with hl2 + wine-git.Here are some examples of approximate average FPS with HL2: wine-1.1.15 = 15-35 fps wine-1.1.15 + remove srgb reload hack = 18-50fps wine-git = 60-100 fps
Left4Dead: wine-1.1.15 = 12-20 fps wine-1.1.15 + remove srgb reload hack = 25-50fps wine-git = 10-22 fps
As you can see, the git patches improve HL2 greatly but stay about the same with left4dead.With the srgb patch + wine 1.1.15 HL2 improves somewhat and left4dead gets a quite decent speed increase (portal has similar results).
The newer patch is obviously the better solution provided that it can be made to work properly with left4dead and portal, as it gives a much greater speed increase.
I've attached a version of the warhammer srgb patch that has the vertex blending portion removed and applies cleanly against wine 1.1.15.All it does is comment out the SRGB reload section.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #27 from DL taedium_vitae@eml.cc 2009-02-24 23:03:39 --- Created an attachment (id=19637) --> (http://bugs.winehq.org/attachment.cgi?id=19637) Removes srgb reloading
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #28 from James Henderson james.m.henderson@gmail.com 2009-02-25 17:08:50 --- I tried with the latest git (wine-1.1.15-391-gb86552f).
The IWineD3DVolumeImpl is now not occuring, but the following line is showing up a lot:
fixme:d3d:state_blendop Unrecognized/Unhandled D3DBLENDOP value 0
Performance is still very poor (10-20 fps).
I tried commenting out the SRGB section as DL did, but it made all textures appear as solid green, even on the menu of the game, I was unable to get into game since everything was green.
http://bugs.winehq.org/show_bug.cgi?id=12453
Juan Battle_Css@hotmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |Battle_Css@hotmail.com
--- Comment #29 from Juan Battle_Css@hotmail.com 2009-03-11 21:20:15 --- I think the srgb reloading patch doesnt work any more in latest version (1.1.16) i think because of the line isnt 130 any more it is 135,
http://bugs.winehq.org/show_bug.cgi?id=12453
Neil Skrypuch ns03ja@brocku.ca changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |ns03ja@brocku.ca
http://bugs.winehq.org/show_bug.cgi?id=12453
G.vorr bugs.glenn+wine@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |bugs.glenn+wine@gmail.com
http://bugs.winehq.org/show_bug.cgi?id=12453
Artem Semendyaev lifeissecret@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |lifeissecret@gmail.com
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #30 from Paul "TBBle" Hampson Paul.Hampson@Pobox.com 2009-05-13 05:03:41 --- Looking at Left4Dead's activity, the main culprit for 'Downloading rgb texture to reload it as srgb' appears to be a texture that matches the window size, reloaded every frame, and with WINED3DUSAGE_RENDERTARGET set.
This suggests to my limited understanding that they're rendering the whole scene out in RGB to a texture, and then redrawing it to the screen in SRGB, which is why this bug has such a large effect on framerate.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #30 from Paul "TBBle" Hampson Paul.Hampson@Pobox.com 2009-05-13 05:03:41 --- Looking at Left4Dead's activity, the main culprit for 'Downloading rgb texture to reload it as srgb' appears to be a texture that matches the window size, reloaded every frame, and with WINED3DUSAGE_RENDERTARGET set.
This suggests to my limited understanding that they're rendering the whole scene out in RGB to a texture, and then redrawing it to the screen in SRGB, which is why this bug has such a large effect on framerate.
--- Comment #31 from Paul "TBBle" Hampson Paul.Hampson@Pobox.com 2009-05-13 05:08:51 --- Created an attachment (id=21072) --> (http://bugs.winehq.org/attachment.cgi?id=21072) Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #30 from Paul "TBBle" Hampson Paul.Hampson@Pobox.com 2009-05-13 05:03:41 --- Looking at Left4Dead's activity, the main culprit for 'Downloading rgb texture to reload it as srgb' appears to be a texture that matches the window size, reloaded every frame, and with WINED3DUSAGE_RENDERTARGET set.
This suggests to my limited understanding that they're rendering the whole scene out in RGB to a texture, and then redrawing it to the screen in SRGB, which is why this bug has such a large effect on framerate.
--- Comment #31 from Paul "TBBle" Hampson Paul.Hampson@Pobox.com 2009-05-13 05:08:51 --- Created an attachment (id=21072) --> (http://bugs.winehq.org/attachment.cgi?id=21072) Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
--- Comment #32 from Tobias Jakobi liquid.acid@gmx.net 2009-05-13 05:52:07 --- I also noticed that. L4D seems to always use a full scene post-processing effect.
On my old Geforce FX 5900 I only get a screen filled with solid color because of this bug: http://bugs.winehq.org/show_bug.cgi?id=14762
So basically L4D renders the scene to a NP2 texture with screen dimensions, then applies various effects (like blood stains) and blits the texture to screen.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #30 from Paul "TBBle" Hampson Paul.Hampson@Pobox.com 2009-05-13 05:03:41 --- Looking at Left4Dead's activity, the main culprit for 'Downloading rgb texture to reload it as srgb' appears to be a texture that matches the window size, reloaded every frame, and with WINED3DUSAGE_RENDERTARGET set.
This suggests to my limited understanding that they're rendering the whole scene out in RGB to a texture, and then redrawing it to the screen in SRGB, which is why this bug has such a large effect on framerate.
--- Comment #31 from Paul "TBBle" Hampson Paul.Hampson@Pobox.com 2009-05-13 05:08:51 --- Created an attachment (id=21072) --> (http://bugs.winehq.org/attachment.cgi?id=21072) Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
--- Comment #32 from Tobias Jakobi liquid.acid@gmx.net 2009-05-13 05:52:07 --- I also noticed that. L4D seems to always use a full scene post-processing effect.
On my old Geforce FX 5900 I only get a screen filled with solid color because of this bug: http://bugs.winehq.org/show_bug.cgi?id=14762
So basically L4D renders the scene to a NP2 texture with screen dimensions, then applies various effects (like blood stains) and blits the texture to screen.
--- Comment #33 from Paul "TBBle" Hampson Paul.Hampson@Pobox.com 2009-05-14 08:30:27 --- (In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
This patch makes Left4Dead noticeably brighter, and noticeably faster.
Note that it turns Warhammer Online completely black, so it's probably not a correct hack.
http://bugs.winehq.org/show_bug.cgi?id=12453
Nicolas Laurenti kleph@kleph.info changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |kleph@kleph.info
--- Comment #30 from Paul "TBBle" Hampson Paul.Hampson@Pobox.com 2009-05-13 05:03:41 --- Looking at Left4Dead's activity, the main culprit for 'Downloading rgb texture to reload it as srgb' appears to be a texture that matches the window size, reloaded every frame, and with WINED3DUSAGE_RENDERTARGET set.
This suggests to my limited understanding that they're rendering the whole scene out in RGB to a texture, and then redrawing it to the screen in SRGB, which is why this bug has such a large effect on framerate.
--- Comment #31 from Paul "TBBle" Hampson Paul.Hampson@Pobox.com 2009-05-13 05:08:51 --- Created an attachment (id=21072) --> (http://bugs.winehq.org/attachment.cgi?id=21072) Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
--- Comment #32 from Tobias Jakobi liquid.acid@gmx.net 2009-05-13 05:52:07 --- I also noticed that. L4D seems to always use a full scene post-processing effect.
On my old Geforce FX 5900 I only get a screen filled with solid color because of this bug: http://bugs.winehq.org/show_bug.cgi?id=14762
So basically L4D renders the scene to a NP2 texture with screen dimensions, then applies various effects (like blood stains) and blits the texture to screen.
--- Comment #33 from Paul "TBBle" Hampson Paul.Hampson@Pobox.com 2009-05-14 08:30:27 --- (In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
This patch makes Left4Dead noticeably brighter, and noticeably faster.
Note that it turns Warhammer Online completely black, so it's probably not a correct hack.
--- Comment #34 from Nicolas Laurenti kleph@kleph.info 2009-05-16 07:44:43 --- (In reply to comment #33)
(In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details] [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
This patch makes Left4Dead noticeably brighter, and noticeably faster.
Note that it turns Warhammer Online completely black, so it's probably not a correct hack.
Warhammer Online is still working here with this patch.
The patch had the expected results on L4D, but I did not notice improvement in Warhammer framerate.
Maybe the patch does help but my config is just a little short: Athlon XP 2800+ 2 Gb of RAM Geforce 68000 /128Mb of VRAM (with the latest nvidia drivers (180.51)
http://bugs.winehq.org/show_bug.cgi?id=12453
Nicolas "kleph" Laurenti kleph@kleph.info changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |kleph@kleph.info
Finite finite999@hotmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |finite999@hotmail.com
--- Comment #30 from Paul "TBBle" Hampson Paul.Hampson@Pobox.com 2009-05-13 05:03:41 --- Looking at Left4Dead's activity, the main culprit for 'Downloading rgb texture to reload it as srgb' appears to be a texture that matches the window size, reloaded every frame, and with WINED3DUSAGE_RENDERTARGET set.
This suggests to my limited understanding that they're rendering the whole scene out in RGB to a texture, and then redrawing it to the screen in SRGB, which is why this bug has such a large effect on framerate.
--- Comment #31 from Paul "TBBle" Hampson Paul.Hampson@Pobox.com 2009-05-13 05:08:51 --- Created an attachment (id=21072) --> (http://bugs.winehq.org/attachment.cgi?id=21072) Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
--- Comment #32 from Tobias Jakobi liquid.acid@gmx.net 2009-05-13 05:52:07 --- I also noticed that. L4D seems to always use a full scene post-processing effect.
On my old Geforce FX 5900 I only get a screen filled with solid color because of this bug: http://bugs.winehq.org/show_bug.cgi?id=14762
So basically L4D renders the scene to a NP2 texture with screen dimensions, then applies various effects (like blood stains) and blits the texture to screen.
--- Comment #33 from Paul "TBBle" Hampson Paul.Hampson@Pobox.com 2009-05-14 08:30:27 --- (In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
This patch makes Left4Dead noticeably brighter, and noticeably faster.
Note that it turns Warhammer Online completely black, so it's probably not a correct hack.
--- Comment #34 from Nicolas "kleph" Laurenti kleph@kleph.info 2009-05-16 07:44:43 --- (In reply to comment #33)
(In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details] [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
This patch makes Left4Dead noticeably brighter, and noticeably faster.
Note that it turns Warhammer Online completely black, so it's probably not a correct hack.
Warhammer Online is still working here with this patch.
The patch had the expected results on L4D, but I did not notice improvement in Warhammer framerate.
Maybe the patch does help but my config is just a little short: Athlon XP 2800+ 2 Gb of RAM Geforce 68000 /128Mb of VRAM (with the latest nvidia drivers (180.51)
--- Comment #35 from Finite finite999@hotmail.com 2009-05-22 00:35:45 --- (In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
This patch works perfectly for Left4Dead single player. However in multiplayer the flashlight no longer works (no projected light effect at all, only the "glow" on the gun is visible). I guess the engine must use different shaders for this effect for online play, and something about this patch breaks that.
http://bugs.winehq.org/show_bug.cgi?id=12453
Nicolas "kleph" Laurenti kleph@kleph.info changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |kleph@kleph.info
Finite finite999@hotmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |finite999@hotmail.com
haarp liquitsnake@gmx.net changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |liquitsnake@gmx.net
--- Comment #30 from Paul "TBBle" Hampson Paul.Hampson@Pobox.com 2009-05-13 05:03:41 --- Looking at Left4Dead's activity, the main culprit for 'Downloading rgb texture to reload it as srgb' appears to be a texture that matches the window size, reloaded every frame, and with WINED3DUSAGE_RENDERTARGET set.
This suggests to my limited understanding that they're rendering the whole scene out in RGB to a texture, and then redrawing it to the screen in SRGB, which is why this bug has such a large effect on framerate.
--- Comment #31 from Paul "TBBle" Hampson Paul.Hampson@Pobox.com 2009-05-13 05:08:51 --- Created an attachment (id=21072) --> (http://bugs.winehq.org/attachment.cgi?id=21072) Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
--- Comment #32 from Tobias Jakobi liquid.acid@gmx.net 2009-05-13 05:52:07 --- I also noticed that. L4D seems to always use a full scene post-processing effect.
On my old Geforce FX 5900 I only get a screen filled with solid color because of this bug: http://bugs.winehq.org/show_bug.cgi?id=14762
So basically L4D renders the scene to a NP2 texture with screen dimensions, then applies various effects (like blood stains) and blits the texture to screen.
--- Comment #33 from Paul "TBBle" Hampson Paul.Hampson@Pobox.com 2009-05-14 08:30:27 --- (In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
This patch makes Left4Dead noticeably brighter, and noticeably faster.
Note that it turns Warhammer Online completely black, so it's probably not a correct hack.
--- Comment #34 from Nicolas "kleph" Laurenti kleph@kleph.info 2009-05-16 07:44:43 --- (In reply to comment #33)
(In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details] [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
This patch makes Left4Dead noticeably brighter, and noticeably faster.
Note that it turns Warhammer Online completely black, so it's probably not a correct hack.
Warhammer Online is still working here with this patch.
The patch had the expected results on L4D, but I did not notice improvement in Warhammer framerate.
Maybe the patch does help but my config is just a little short: Athlon XP 2800+ 2 Gb of RAM Geforce 68000 /128Mb of VRAM (with the latest nvidia drivers (180.51)
--- Comment #35 from Finite finite999@hotmail.com 2009-05-22 00:35:45 --- (In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
This patch works perfectly for Left4Dead single player. However in multiplayer the flashlight no longer works (no projected light effect at all, only the "glow" on the gun is visible). I guess the engine must use different shaders for this effect for online play, and something about this patch breaks that.
http://bugs.winehq.org/show_bug.cgi?id=12453
Alexandre Lima formulaire@bacatta.name changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |formulaire@bacatta.name
http://bugs.winehq.org/show_bug.cgi?id=12453
brooss.teambb@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |brooss.teambb@gmail.com
http://bugs.winehq.org/show_bug.cgi?id=12453
donotreply@who.net changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |donotreply@who.net
--- Comment #36 from donotreply@who.net 2009-06-30 09:49:57 --- (In reply to comment #35)
(In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details] [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
This patch works perfectly for Left4Dead single player. However in multiplayer the flashlight no longer works (no projected light effect at all, only the "glow" on the gun is visible). I guess the engine must use different shaders for this effect for online play, and something about this patch breaks that.
The patch works perfectly for me against 1.1.24. Flashlight works as intended in multiplayer campaign. I would like to see this get implemented as the FPS increases are great. I no longer have to disable HDR in Left 4 Dead, albeit settings must still be on low for models, textures, etc. and the resolution doesn't go past 960x600. I know that sounds preachy and stuff but this patch is a step in the right direction.
http://bugs.winehq.org/show_bug.cgi?id=12453
donotreply@who.net changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |donotreply@who.net
Kari refic@psimerion.org changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |refic@psimerion.org
--- Comment #36 from donotreply@who.net 2009-06-30 09:49:57 --- (In reply to comment #35)
(In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details] [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
This patch works perfectly for Left4Dead single player. However in multiplayer the flashlight no longer works (no projected light effect at all, only the "glow" on the gun is visible). I guess the engine must use different shaders for this effect for online play, and something about this patch breaks that.
The patch works perfectly for me against 1.1.24. Flashlight works as intended in multiplayer campaign. I would like to see this get implemented as the FPS increases are great. I no longer have to disable HDR in Left 4 Dead, albeit settings must still be on low for models, textures, etc. and the resolution doesn't go past 960x600. I know that sounds preachy and stuff but this patch is a step in the right direction.
--- Comment #37 from Kari refic@psimerion.org 2009-06-30 10:30:04 --- Patch working extremely well here on 1.1.24. I got over double FPS (70-80) most of the time!
http://bugs.winehq.org/show_bug.cgi?id=12453
donotreply@who.net changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |donotreply@who.net
Kari refic@psimerion.org changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |refic@psimerion.org
--- Comment #36 from donotreply@who.net 2009-06-30 09:49:57 --- (In reply to comment #35)
(In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details] [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
This patch works perfectly for Left4Dead single player. However in multiplayer the flashlight no longer works (no projected light effect at all, only the "glow" on the gun is visible). I guess the engine must use different shaders for this effect for online play, and something about this patch breaks that.
The patch works perfectly for me against 1.1.24. Flashlight works as intended in multiplayer campaign. I would like to see this get implemented as the FPS increases are great. I no longer have to disable HDR in Left 4 Dead, albeit settings must still be on low for models, textures, etc. and the resolution doesn't go past 960x600. I know that sounds preachy and stuff but this patch is a step in the right direction.
--- Comment #37 from Kari refic@psimerion.org 2009-06-30 10:30:04 --- Patch working extremely well here on 1.1.24. I got over double FPS (70-80) most of the time!
--- Comment #38 from donotreply@who.net 2009-07-16 03:25:47 --- I can in fact raise the eyecandy with this patch to respectable levels (medium shaders, models/text. high, effects high) ALTHOUGH regular shadows are broken for characters, zombies, and etc. Other than that its a HUGE performance boost.
http://bugs.winehq.org/show_bug.cgi?id=12453
donotreply@who.net changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |donotreply@who.net
Kari refic@psimerion.org changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |refic@psimerion.org
--- Comment #36 from donotreply@who.net 2009-06-30 09:49:57 --- (In reply to comment #35)
(In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details] [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
This patch works perfectly for Left4Dead single player. However in multiplayer the flashlight no longer works (no projected light effect at all, only the "glow" on the gun is visible). I guess the engine must use different shaders for this effect for online play, and something about this patch breaks that.
The patch works perfectly for me against 1.1.24. Flashlight works as intended in multiplayer campaign. I would like to see this get implemented as the FPS increases are great. I no longer have to disable HDR in Left 4 Dead, albeit settings must still be on low for models, textures, etc. and the resolution doesn't go past 960x600. I know that sounds preachy and stuff but this patch is a step in the right direction.
--- Comment #37 from Kari refic@psimerion.org 2009-06-30 10:30:04 --- Patch working extremely well here on 1.1.24. I got over double FPS (70-80) most of the time!
--- Comment #38 from donotreply@who.net 2009-07-16 03:25:47 --- I can in fact raise the eyecandy with this patch to respectable levels (medium shaders, models/text. high, effects high) ALTHOUGH regular shadows are broken for characters, zombies, and etc. Other than that its a HUGE performance boost.
--- Comment #39 from donotreply@who.net 2009-08-02 06:59:07 --- (In reply to comment #35)
(In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details] [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
This patch works perfectly for Left4Dead single player. However in multiplayer the flashlight no longer works (no projected light effect at all, only the "glow" on the gun is visible). I guess the engine must use different shaders for this effect for online play, and something about this patch breaks that.
Huh, another update, funnily enough I get this corruption now against 1.1.26. Although this coupled with using ARB makes L4D run smooth enough. Since I am unable to downgrade from 1.1.26 to 1.1.24 (the wine folder gets "corrupted" when updated to 1.1.26 and steam crashes on launch when you revert back) I would wish that this error gets worked out.
http://bugs.winehq.org/show_bug.cgi?id=12453
donotreply@who.net changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |donotreply@who.net
Kari refic@psimerion.org changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |refic@psimerion.org
--- Comment #36 from donotreply@who.net 2009-06-30 09:49:57 --- (In reply to comment #35)
(In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details] [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
This patch works perfectly for Left4Dead single player. However in multiplayer the flashlight no longer works (no projected light effect at all, only the "glow" on the gun is visible). I guess the engine must use different shaders for this effect for online play, and something about this patch breaks that.
The patch works perfectly for me against 1.1.24. Flashlight works as intended in multiplayer campaign. I would like to see this get implemented as the FPS increases are great. I no longer have to disable HDR in Left 4 Dead, albeit settings must still be on low for models, textures, etc. and the resolution doesn't go past 960x600. I know that sounds preachy and stuff but this patch is a step in the right direction.
--- Comment #37 from Kari refic@psimerion.org 2009-06-30 10:30:04 --- Patch working extremely well here on 1.1.24. I got over double FPS (70-80) most of the time!
--- Comment #38 from donotreply@who.net 2009-07-16 03:25:47 --- I can in fact raise the eyecandy with this patch to respectable levels (medium shaders, models/text. high, effects high) ALTHOUGH regular shadows are broken for characters, zombies, and etc. Other than that its a HUGE performance boost.
--- Comment #39 from donotreply@who.net 2009-08-02 06:59:07 --- (In reply to comment #35)
(In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details] [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
This patch works perfectly for Left4Dead single player. However in multiplayer the flashlight no longer works (no projected light effect at all, only the "glow" on the gun is visible). I guess the engine must use different shaders for this effect for online play, and something about this patch breaks that.
Huh, another update, funnily enough I get this corruption now against 1.1.26. Although this coupled with using ARB makes L4D run smooth enough. Since I am unable to downgrade from 1.1.26 to 1.1.24 (the wine folder gets "corrupted" when updated to 1.1.26 and steam crashes on launch when you revert back) I would wish that this error gets worked out.
--- Comment #40 from Chris Boyle chris@boyle.name 2009-08-06 17:24:35 --- To resolve the confusion between comment 35 and comment 36: I find in wine 1.1.26~winehq0~ubuntu~9.04-0ubuntu1 both with and without the patch, the flashlight is fine in all modes except versus. So, multiplayer campaign is fine, multiplayer versus is not: no flashlight, and no flashlights from other players either.
http://bugs.winehq.org/show_bug.cgi?id=12453
donotreply@who.net changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |donotreply@who.net
Kari refic@psimerion.org changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |refic@psimerion.org
Jesse A. ep@trishock.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |ep@trishock.com
--- Comment #36 from donotreply@who.net 2009-06-30 09:49:57 --- (In reply to comment #35)
(In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details] [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
This patch works perfectly for Left4Dead single player. However in multiplayer the flashlight no longer works (no projected light effect at all, only the "glow" on the gun is visible). I guess the engine must use different shaders for this effect for online play, and something about this patch breaks that.
The patch works perfectly for me against 1.1.24. Flashlight works as intended in multiplayer campaign. I would like to see this get implemented as the FPS increases are great. I no longer have to disable HDR in Left 4 Dead, albeit settings must still be on low for models, textures, etc. and the resolution doesn't go past 960x600. I know that sounds preachy and stuff but this patch is a step in the right direction.
--- Comment #37 from Kari refic@psimerion.org 2009-06-30 10:30:04 --- Patch working extremely well here on 1.1.24. I got over double FPS (70-80) most of the time!
--- Comment #38 from donotreply@who.net 2009-07-16 03:25:47 --- I can in fact raise the eyecandy with this patch to respectable levels (medium shaders, models/text. high, effects high) ALTHOUGH regular shadows are broken for characters, zombies, and etc. Other than that its a HUGE performance boost.
--- Comment #39 from donotreply@who.net 2009-08-02 06:59:07 --- (In reply to comment #35)
(In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details] [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
This patch works perfectly for Left4Dead single player. However in multiplayer the flashlight no longer works (no projected light effect at all, only the "glow" on the gun is visible). I guess the engine must use different shaders for this effect for online play, and something about this patch breaks that.
Huh, another update, funnily enough I get this corruption now against 1.1.26. Although this coupled with using ARB makes L4D run smooth enough. Since I am unable to downgrade from 1.1.26 to 1.1.24 (the wine folder gets "corrupted" when updated to 1.1.26 and steam crashes on launch when you revert back) I would wish that this error gets worked out.
--- Comment #40 from Chris Boyle chris@boyle.name 2009-08-06 17:24:35 --- To resolve the confusion between comment 35 and comment 36: I find in wine 1.1.26~winehq0~ubuntu~9.04-0ubuntu1 both with and without the patch, the flashlight is fine in all modes except versus. So, multiplayer campaign is fine, multiplayer versus is not: no flashlight, and no flashlights from other players either.
--- Comment #41 from Jesse A. ep@trishock.com 2009-08-08 13:47:13 --- The following references Left 4 Dead:
Using the "Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21" patch on 1.1.27 yielded a performance increase over 1.1.26 without the patch. In multiplayer the flashlight does not work upon freshly loading. Causing the game to lose focus and then re-focusing it causes the flashlight to work properly again (Alt-tab out, change workspaces, etc). I made a more lengthy post on the Left 4 Dead page on appdb but I will post the results here also:
With all settings on low (except trilinear filtering enabled) at 960x540 I made a timedemo and ran it on 1.1.26 and 1.1.27 with the above patch. On 1.1.26 the timedemo reported 31.02 fps and on the patched 1.1.27 the timedemo reported 40.56 fps.
Not exactly a monumental increase in framerate but definitely a big step in the right direction.
http://bugs.winehq.org/show_bug.cgi?id=12453
donotreply@who.net changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |donotreply@who.net
Kari refic@psimerion.org changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |refic@psimerion.org
Jesse A. ep@trishock.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |ep@trishock.com
--- Comment #36 from donotreply@who.net 2009-06-30 09:49:57 --- (In reply to comment #35)
(In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details] [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
This patch works perfectly for Left4Dead single player. However in multiplayer the flashlight no longer works (no projected light effect at all, only the "glow" on the gun is visible). I guess the engine must use different shaders for this effect for online play, and something about this patch breaks that.
The patch works perfectly for me against 1.1.24. Flashlight works as intended in multiplayer campaign. I would like to see this get implemented as the FPS increases are great. I no longer have to disable HDR in Left 4 Dead, albeit settings must still be on low for models, textures, etc. and the resolution doesn't go past 960x600. I know that sounds preachy and stuff but this patch is a step in the right direction.
--- Comment #37 from Kari refic@psimerion.org 2009-06-30 10:30:04 --- Patch working extremely well here on 1.1.24. I got over double FPS (70-80) most of the time!
--- Comment #38 from donotreply@who.net 2009-07-16 03:25:47 --- I can in fact raise the eyecandy with this patch to respectable levels (medium shaders, models/text. high, effects high) ALTHOUGH regular shadows are broken for characters, zombies, and etc. Other than that its a HUGE performance boost.
--- Comment #39 from donotreply@who.net 2009-08-02 06:59:07 --- (In reply to comment #35)
(In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details] [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
This patch works perfectly for Left4Dead single player. However in multiplayer the flashlight no longer works (no projected light effect at all, only the "glow" on the gun is visible). I guess the engine must use different shaders for this effect for online play, and something about this patch breaks that.
Huh, another update, funnily enough I get this corruption now against 1.1.26. Although this coupled with using ARB makes L4D run smooth enough. Since I am unable to downgrade from 1.1.26 to 1.1.24 (the wine folder gets "corrupted" when updated to 1.1.26 and steam crashes on launch when you revert back) I would wish that this error gets worked out.
--- Comment #40 from Chris Boyle chris@boyle.name 2009-08-06 17:24:35 --- To resolve the confusion between comment 35 and comment 36: I find in wine 1.1.26~winehq0~ubuntu~9.04-0ubuntu1 both with and without the patch, the flashlight is fine in all modes except versus. So, multiplayer campaign is fine, multiplayer versus is not: no flashlight, and no flashlights from other players either.
--- Comment #41 from Jesse A. ep@trishock.com 2009-08-08 13:47:13 --- The following references Left 4 Dead:
Using the "Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21" patch on 1.1.27 yielded a performance increase over 1.1.26 without the patch. In multiplayer the flashlight does not work upon freshly loading. Causing the game to lose focus and then re-focusing it causes the flashlight to work properly again (Alt-tab out, change workspaces, etc). I made a more lengthy post on the Left 4 Dead page on appdb but I will post the results here also:
With all settings on low (except trilinear filtering enabled) at 960x540 I made a timedemo and ran it on 1.1.26 and 1.1.27 with the above patch. On 1.1.26 the timedemo reported 31.02 fps and on the patched 1.1.27 the timedemo reported 40.56 fps.
Not exactly a monumental increase in framerate but definitely a big step in the right direction.
--- Comment #42 from donotreply@who.net 2009-08-22 18:24:54 --- (In reply to comment #41)
The following references Left 4 Dead:
Using the "Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21" patch on 1.1.27 yielded a performance increase over 1.1.26 without the patch. In multiplayer the flashlight does not work upon freshly loading. Causing the game to lose focus and then re-focusing it causes the flashlight to work properly again (Alt-tab out, change workspaces, etc). I made a more lengthy post on the Left 4 Dead page on appdb but I will post the results here also:
With all settings on low (except trilinear filtering enabled) at 960x540 I made a timedemo and ran it on 1.1.26 and 1.1.27 with the above patch. On 1.1.26 the timedemo reported 31.02 fps and on the patched 1.1.27 the timedemo reported 40.56 fps.
Not exactly a monumental increase in framerate but definitely a big step in the right direction.
Something after 1.1.24 makes this patch poo-poo. I reverted back in a new prefix using this patch and I forget how awesome the performance was. Performance gains don't seem to be there as much using 1.1.25+.
http://bugs.winehq.org/show_bug.cgi?id=12453
donotreply@who.net changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |donotreply@who.net
Kari refic@psimerion.org changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |refic@psimerion.org
Jesse A. ep@trishock.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |ep@trishock.com
Laissus Yoann yoann.laissus@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |yoann.laissus@gmail.com
--- Comment #36 from donotreply@who.net 2009-06-30 09:49:57 --- (In reply to comment #35)
(In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details] [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
This patch works perfectly for Left4Dead single player. However in multiplayer the flashlight no longer works (no projected light effect at all, only the "glow" on the gun is visible). I guess the engine must use different shaders for this effect for online play, and something about this patch breaks that.
The patch works perfectly for me against 1.1.24. Flashlight works as intended in multiplayer campaign. I would like to see this get implemented as the FPS increases are great. I no longer have to disable HDR in Left 4 Dead, albeit settings must still be on low for models, textures, etc. and the resolution doesn't go past 960x600. I know that sounds preachy and stuff but this patch is a step in the right direction.
--- Comment #37 from Kari refic@psimerion.org 2009-06-30 10:30:04 --- Patch working extremely well here on 1.1.24. I got over double FPS (70-80) most of the time!
--- Comment #38 from donotreply@who.net 2009-07-16 03:25:47 --- I can in fact raise the eyecandy with this patch to respectable levels (medium shaders, models/text. high, effects high) ALTHOUGH regular shadows are broken for characters, zombies, and etc. Other than that its a HUGE performance boost.
--- Comment #39 from donotreply@who.net 2009-08-02 06:59:07 --- (In reply to comment #35)
(In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details] [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
This patch works perfectly for Left4Dead single player. However in multiplayer the flashlight no longer works (no projected light effect at all, only the "glow" on the gun is visible). I guess the engine must use different shaders for this effect for online play, and something about this patch breaks that.
Huh, another update, funnily enough I get this corruption now against 1.1.26. Although this coupled with using ARB makes L4D run smooth enough. Since I am unable to downgrade from 1.1.26 to 1.1.24 (the wine folder gets "corrupted" when updated to 1.1.26 and steam crashes on launch when you revert back) I would wish that this error gets worked out.
--- Comment #40 from Chris Boyle chris@boyle.name 2009-08-06 17:24:35 --- To resolve the confusion between comment 35 and comment 36: I find in wine 1.1.26~winehq0~ubuntu~9.04-0ubuntu1 both with and without the patch, the flashlight is fine in all modes except versus. So, multiplayer campaign is fine, multiplayer versus is not: no flashlight, and no flashlights from other players either.
--- Comment #41 from Jesse A. ep@trishock.com 2009-08-08 13:47:13 --- The following references Left 4 Dead:
Using the "Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21" patch on 1.1.27 yielded a performance increase over 1.1.26 without the patch. In multiplayer the flashlight does not work upon freshly loading. Causing the game to lose focus and then re-focusing it causes the flashlight to work properly again (Alt-tab out, change workspaces, etc). I made a more lengthy post on the Left 4 Dead page on appdb but I will post the results here also:
With all settings on low (except trilinear filtering enabled) at 960x540 I made a timedemo and ran it on 1.1.26 and 1.1.27 with the above patch. On 1.1.26 the timedemo reported 31.02 fps and on the patched 1.1.27 the timedemo reported 40.56 fps.
Not exactly a monumental increase in framerate but definitely a big step in the right direction.
--- Comment #42 from donotreply@who.net 2009-08-22 18:24:54 --- (In reply to comment #41)
The following references Left 4 Dead:
Using the "Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21" patch on 1.1.27 yielded a performance increase over 1.1.26 without the patch. In multiplayer the flashlight does not work upon freshly loading. Causing the game to lose focus and then re-focusing it causes the flashlight to work properly again (Alt-tab out, change workspaces, etc). I made a more lengthy post on the Left 4 Dead page on appdb but I will post the results here also:
With all settings on low (except trilinear filtering enabled) at 960x540 I made a timedemo and ran it on 1.1.26 and 1.1.27 with the above patch. On 1.1.26 the timedemo reported 31.02 fps and on the patched 1.1.27 the timedemo reported 40.56 fps.
Not exactly a monumental increase in framerate but definitely a big step in the right direction.
Something after 1.1.24 makes this patch poo-poo. I reverted back in a new prefix using this patch and I forget how awesome the performance was. Performance gains don't seem to be there as much using 1.1.25+.
http://bugs.winehq.org/show_bug.cgi?id=12453
donotreply@who.net changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |donotreply@who.net
Kari refic@psimerion.org changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |refic@psimerion.org
Jesse A. ep@trishock.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |ep@trishock.com
Laissus Yoann yoann.laissus@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |yoann.laissus@gmail.com
msn@gaiatools.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |msn@gaiatools.com
--- Comment #36 from donotreply@who.net 2009-06-30 09:49:57 --- (In reply to comment #35)
(In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details] [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
This patch works perfectly for Left4Dead single player. However in multiplayer the flashlight no longer works (no projected light effect at all, only the "glow" on the gun is visible). I guess the engine must use different shaders for this effect for online play, and something about this patch breaks that.
The patch works perfectly for me against 1.1.24. Flashlight works as intended in multiplayer campaign. I would like to see this get implemented as the FPS increases are great. I no longer have to disable HDR in Left 4 Dead, albeit settings must still be on low for models, textures, etc. and the resolution doesn't go past 960x600. I know that sounds preachy and stuff but this patch is a step in the right direction.
--- Comment #37 from Kari refic@psimerion.org 2009-06-30 10:30:04 --- Patch working extremely well here on 1.1.24. I got over double FPS (70-80) most of the time!
--- Comment #38 from donotreply@who.net 2009-07-16 03:25:47 --- I can in fact raise the eyecandy with this patch to respectable levels (medium shaders, models/text. high, effects high) ALTHOUGH regular shadows are broken for characters, zombies, and etc. Other than that its a HUGE performance boost.
--- Comment #39 from donotreply@who.net 2009-08-02 06:59:07 --- (In reply to comment #35)
(In reply to comment #31)
Created an attachment (id=21072)
--> (http://bugs.winehq.org/attachment.cgi?id=21072) [details] [details]
Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21
Here's a hack patch against 1.1.21 or so which should bypass attempts to reload an RGB texture as an SRGB texture or vice-versa.
Depending on how the texture was first uploaded, it'll either be too bright (trying to reload RGB as SRGB) or too dark (trying to reload SRGB as RGB).
This patch makes Left4Dead noticeably brighter, and noticeably faster.
This patch works perfectly for Left4Dead single player. However in multiplayer the flashlight no longer works (no projected light effect at all, only the "glow" on the gun is visible). I guess the engine must use different shaders for this effect for online play, and something about this patch breaks that.
Huh, another update, funnily enough I get this corruption now against 1.1.26. Although this coupled with using ARB makes L4D run smooth enough. Since I am unable to downgrade from 1.1.26 to 1.1.24 (the wine folder gets "corrupted" when updated to 1.1.26 and steam crashes on launch when you revert back) I would wish that this error gets worked out.
--- Comment #40 from Chris Boyle chris@boyle.name 2009-08-06 17:24:35 --- To resolve the confusion between comment 35 and comment 36: I find in wine 1.1.26~winehq0~ubuntu~9.04-0ubuntu1 both with and without the patch, the flashlight is fine in all modes except versus. So, multiplayer campaign is fine, multiplayer versus is not: no flashlight, and no flashlights from other players either.
--- Comment #41 from Jesse A. ep@trishock.com 2009-08-08 13:47:13 --- The following references Left 4 Dead:
Using the "Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21" patch on 1.1.27 yielded a performance increase over 1.1.26 without the patch. In multiplayer the flashlight does not work upon freshly loading. Causing the game to lose focus and then re-focusing it causes the flashlight to work properly again (Alt-tab out, change workspaces, etc). I made a more lengthy post on the Left 4 Dead page on appdb but I will post the results here also:
With all settings on low (except trilinear filtering enabled) at 960x540 I made a timedemo and ran it on 1.1.26 and 1.1.27 with the above patch. On 1.1.26 the timedemo reported 31.02 fps and on the patched 1.1.27 the timedemo reported 40.56 fps.
Not exactly a monumental increase in framerate but definitely a big step in the right direction.
--- Comment #42 from donotreply@who.net 2009-08-22 18:24:54 --- (In reply to comment #41)
The following references Left 4 Dead:
Using the "Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.21" patch on 1.1.27 yielded a performance increase over 1.1.26 without the patch. In multiplayer the flashlight does not work upon freshly loading. Causing the game to lose focus and then re-focusing it causes the flashlight to work properly again (Alt-tab out, change workspaces, etc). I made a more lengthy post on the Left 4 Dead page on appdb but I will post the results here also:
With all settings on low (except trilinear filtering enabled) at 960x540 I made a timedemo and ran it on 1.1.26 and 1.1.27 with the above patch. On 1.1.26 the timedemo reported 31.02 fps and on the patched 1.1.27 the timedemo reported 40.56 fps.
Not exactly a monumental increase in framerate but definitely a big step in the right direction.
Something after 1.1.24 makes this patch poo-poo. I reverted back in a new prefix using this patch and I forget how awesome the performance was. Performance gains don't seem to be there as much using 1.1.25+.
http://bugs.winehq.org/show_bug.cgi?id=12453
Vitaliy Margolen vitaliy@kievinfo.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED
--- Comment #43 from Vitaliy Margolen vitaliy@kievinfo.com 2009-10-29 00:03:42 --- SRGB slowdown been fixed for a while. All new slowdowns have a different source.
http://bugs.winehq.org/show_bug.cgi?id=12453
Luke Bratch l_bratch@yahoo.co.uk changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |l_bratch@yahoo.co.uk
http://bugs.winehq.org/show_bug.cgi?id=12453
Sven-Hendrik Haase sh@lutzhaase.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |sh@lutzhaase.com
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #44 from Sven-Hendrik Haase sh@lutzhaase.com 2009-11-09 17:33:56 --- Reading the comments on http://appdb.winehq.org/objectManager.php?sClass=version&iId=9901 it appears this bug is back with wine 1.1.31. Shall we reopen this or clone it?
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #45 from Vitaliy Margolen vitaliy@kievinfo.com 2009-11-09 21:43:23 --- (In reply to comment #44) No, there's already a different bug for that.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #46 from Sven-Hendrik Haase sh@lutzhaase.com 2009-11-10 08:53:53 --- (In reply to comment #45)
(In reply to comment #44) No, there's already a different bug for that.
Can you point me to that bug? It doesn't appear to be linked to the source games.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #47 from Vitaliy Margolen vitaliy@kievinfo.com 2009-11-10 20:32:05 --- (In reply to comment #46) For example bug 20602. If you want to double check - do a regression testing.
This bug was about SRGB causing textures to be reloaded all the time. This was fixed a while ago.
http://bugs.winehq.org/show_bug.cgi?id=12453
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #48 from Alexandre Julliard julliard@winehq.org 2009-11-13 12:42:33 --- Closing bugs fixed in 1.1.33.
http://bugs.winehq.org/show_bug.cgi?id=12453
Pavel Valach valach.pavel@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |valach.pavel@gmail.com
--- Comment #49 from Pavel Valach valach.pavel@gmail.com 2009-11-14 15:38:52 --- I am sorry to say this bug wasn't fixed... I'm getting bad slowdowns again in Portal and TF2 with Wine 1.1.33. Launching through Steam.
fixme:d3d_surface:IWineD3DSurfaceImpl_LoadLocation 0x1face0: Downloading rgb texture to reload it as srgb fixme:d3d_surface:IWineD3DSurfaceImpl_LoadLocation 0x15afb360: Downloading rgb texture to reload it as srgb fixme:d3d_surface:IWineD3DSurfaceImpl_LoadLocation 0x14440398: Downloading rgb texture to reload it as srgb
Ubuntu 9.10 x86 nVidia GeForce 8600GT (driver 190.42)
Can you please confirm my results?
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #50 from Laissus Yoann yoann.laissus@gmail.com 2009-11-14 15:53:05 --- (In reply to comment #49)
I am sorry to say this bug wasn't fixed... I'm getting bad slowdowns again in Portal and TF2 with Wine 1.1.33. Launching through Steam.
fixme:d3d_surface:IWineD3DSurfaceImpl_LoadLocation 0x1face0: Downloading rgb texture to reload it as srgb fixme:d3d_surface:IWineD3DSurfaceImpl_LoadLocation 0x15afb360: Downloading rgb texture to reload it as srgb fixme:d3d_surface:IWineD3DSurfaceImpl_LoadLocation 0x14440398: Downloading rgb texture to reload it as srgb
Ubuntu 9.10 x86 nVidia GeForce 8600GT (driver 190.42)
Can you please confirm my results?
I confirm this with 1.1.33 on left 4 dead (same srgb reloading) I have the same framerate as wine 1.1.32 with and without the srgb hack.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #51 from Tobias Jakobi liquid.acid@gmx.net 2009-11-14 17:50:18 --- (In reply to comment #47)
This bug was about SRGB causing textures to be reloaded all the time. This was fixed a while ago.
Can you point me to the commit that supposedly fixed this?
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #52 from James Henderson james.m.henderson@gmail.com 2009-11-14 18:47:27 --- I will also confirm this for wine-1.1.33, just built from git an hour or so ago. fixme:d3d_surface:IWineD3DSurfaceImpl_LoadLocation 0x1041d9b8: Downloading rgb texture to reload it as srgb
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #53 from Paul "TBBle" Hampson Paul.Hampson@Pobox.com 2009-11-14 21:30:08 --- (In reply to comment #51)
(In reply to comment #47)
This bug was about SRGB causing textures to be reloaded all the time. This was fixed a while ago.
Can you point me to the commit that supposedly fixed this?
I'm guessing f3b39d908523ad3641de95e475402ae1436bb4ce which was part of release 1.1.27.
I very very quick skim suggests this change fixes textures that actually start SRGB, to not be reuploaded unneccesarily. It wouldn't have fixed the Warhammer Online issue which was the texture's sampler state being flipped back and forth, triggering reuploads every time, due to difference in Direct3D and OpenGL SRGB support.
http://source.winehq.org/git/wine.git/?a=commitdiff;h=f3b39d908523ad3641de95...
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #54 from Vitaliy Margolen vitaliy@kievinfo.com 2009-11-14 23:13:34 --- (In reply to comment #49)
I am sorry to say this bug wasn't fixed...
I've already said this
THIS BUG IS FIXED (or at least WAS FIXED).
Open new bugs if you have any problems, that might look the same.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #55 from Tobias Jakobi liquid.acid@gmx.net 2009-11-15 02:39:14 --- (In reply to comment #54)
THIS BUG IS FIXED (or at least WAS FIXED).
If it was fixed, then state the commit that fixed it. I don't really see where this was done at all, but maybe I'm just looking at the wrong places.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #56 from Artem S. Tashkinov t.artem@mailcity.com 2009-11-15 03:44:17 --- Vitaliy, if this bug is really fixed then messages like
"fixme:d3d_surface:IWineD3DSurfaceImpl_LoadLocation 0x11f672c8: Downloading rgb texture to reload it as srgb"
are supposed to be gone, right?
Because now I'm running L4D1 under Wine 1.33 and it's still terribly slow (even with all settings set to the lowest values). However I need to say that I get thousands of messages like this one:
fixme:d3d:state_blendop Unrecognized/Unhandled D3DBLENDOP value 0
http://bugs.winehq.org/show_bug.cgi?id=12453
Vitaliy Margolen vitaliy@kievinfo.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|CLOSED |REOPENED Resolution|FIXED |
--- Comment #57 from Vitaliy Margolen vitaliy@kievinfo.com 2009-11-15 11:03:43 --- Fine
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #58 from Night Nord NightNord@gmail.com 2009-11-15 11:24:54 --- Probably someone, who are using git, and knows what last version was working, could do regression testing?
This is an regression, cause it _was_ fixed, really. And regression, imo, should be posted as separate bug tagged as regression.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #59 from Artem S. Tashkinov t.artem@mailcity.com 2009-11-15 14:56:11 --- The only relevant SRGB that I found (http://source.winehq.org/git/wine.git/?a=search&h=HEAD&st=commit&...) was this one:
http://source.winehq.org/git/wine.git?a=commitdiff;h=33ca3c02076082131c0b846...
That means that this version of Wine: http://source.winehq.org/git/wine.git?a=tree;h=82322e025dafc9a190d11a52a6130... should work just fine or this bug has never been fixed in the first place.
If the aforementioned Wine version indeed works (please, someone test it) then we need to find out what Wine git commit has broken the Source Engine.
However I can be all wrong, and the right SRGB fix was another one.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #60 from DL taedium_vitae@eml.cc 2009-11-15 19:19:52 --- The SRGB patch(s) only ever fixed the original Half Life 2.It didn't help with HL2 : EP1/EP2, Portal, Left4Dead or Zeno Clash.It seems that the patch only worked properly with early versions of the source engine, and that it does not work correctly with later versions.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #61 from Kari refic@psimerion.org 2009-11-16 02:08:46 --- (In reply to comment #60)
The SRGB patch(s) only ever fixed the original Half Life 2.It didn't help with HL2 : EP1/EP2, Portal, Left4Dead or Zeno Clash.It seems that the patch only worked properly with early versions of the source engine, and that it does not work correctly with later versions.
Actually the patch that was attached here (#31) did have a nice effect on L4D, FPS doubled after it. For me at least.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #62 from DL taedium_vitae@eml.cc 2009-11-16 02:22:51 --- (In reply to comment #61)
(In reply to comment #60)
The SRGB patch(s) only ever fixed the original Half Life 2.It didn't help with HL2 : EP1/EP2, Portal, Left4Dead or Zeno Clash.It seems that the patch only worked properly with early versions of the source engine, and that it does not work correctly with later versions.
Actually the patch that was attached here (#31) did have a nice effect on L4D, FPS doubled after it. For me at least.
Yes, but that patch is more of a hack than a real fix, and it breaks lighting effects in most games.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #63 from msn@gaiatools.com 2009-11-17 15:36:45 --- Another performance factor in Source games is sound, it seems. Without any sound plugin, TF2 runs well on dxlevel 81 all high at ~15~35 fps (still too slow to be playable), but with the same settings and with the ALSA audio driver, I get 8~15 fps. I also get 10~20 fps with winepulse.
http://bugs.winehq.org/show_bug.cgi?id=12453
Ciaran 15094014@sun.ac.za changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |15094014@sun.ac.za
--- Comment #64 from Ciaran 15094014@sun.ac.za 2009-11-17 18:52:14 --- Tested with wine 1.1.21 and disable srgb sampler patch, it does give an fps boast. Not double but definitely noticeable but makes the game very dark, almost unplayable.
Perhaps is someone could make the same patch for wine 1.1.33 we could check the performance difference and see if that is what causes the slowdown
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #65 from Luke Bratch l_bratch@yahoo.co.uk 2009-11-17 19:03:30 --- Created an attachment (id=24812) --> (http://bugs.winehq.org/attachment.cgi?id=24812) Disable SRGB-sampler-tate-triggered texture reloads, against 1.1.33
Patch against wine-1.1.33.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #66 from msn@gaiatools.com 2009-11-17 21:34:13 --- I can confirm a noticeable FPS boost, but the water looks like ink (and the screen turns all black when you go in water) and I can see lights through walls.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #67 from Ciaran 15094014@sun.ac.za 2009-11-17 23:25:12 --- confirm noticeable fps boast, almost double. Can actually up graphics now, didn't check. Although now for some reason the intro movies were slightly pixelated and laggy.
Has anyone checked for performance boosts in other source games? I currently don't have any others to test.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #68 from donotreply@who.net 2009-11-19 20:10:46 --- (In reply to comment #67)
confirm noticeable fps boast, almost double. Can actually up graphics now, didn't check. Although now for some reason the intro movies were slightly pixelated and laggy.
Has anyone checked for performance boosts in other source games? I currently don't have any others to test.
Small FPS boost, missing lights, and darkened screen in L4D. Pretty much the same deal applying he old patch post 1.1.24. Aslo, there is no performance increase in any other Steam game on my end (HL2, TF2, CS, Garry's Mod, Portal, and a few more).
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #69 from Luke Bratch l_bratch@yahoo.co.uk 2010-02-20 13:31:38 --- This seems to affect the recently released Alien vs Predator demo too.
http://bugs.winehq.org/show_bug.cgi?id=12453
MCpaul34 MCpaul34@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |MCpaul34@gmail.com
--- Comment #70 from MCpaul34 MCpaul34@gmail.com 2010-03-07 07:31:16 --- It's better since wine 1.1.39 (but it's really really far from good conditions) but now there is some render problem (black texture and messed up light effects).
http://bugs.winehq.org/show_bug.cgi?id=12453
Enrico Enrico_m@gmx.de changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |Enrico_m@gmx.de
http://bugs.winehq.org/show_bug.cgi?id=12453
Marcus Hambraeus marcus.hambraeus@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |marcus.hambraeus@gmail.com
--- Comment #71 from Marcus Hambraeus marcus.hambraeus@gmail.com 2010-06-12 02:23:56 --- I can confirm really low fps on two different machines, both running wine 1.2rc2. Tried with several Source-based games (L4D, L4D2, Portal). While playing L4D2, the console gets spammed with messages saying:
fixme:d3d_surface:IWineD3DSurfaceImpl_LoadLocation 0x1fbd0ee0: Downloading rgb texture to reload it as srgb
Is there a patch like for 1.1.21/33 for 1.2rc2 or is there some other magic to do to make this work?
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #72 from Yoann Laissus yoann.laissus@gmail.com 2010-06-12 02:41:39 --- Created an attachment (id=28759) --> (http://bugs.winehq.org/attachment.cgi?id=28759) An updated patch
http://bugs.winehq.org/show_bug.cgi?id=12453
Ken Phillis Jr kphillisjr@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |kphillisjr@gmail.com
--- Comment #73 from Ken Phillis Jr kphillisjr@gmail.com 2010-07-20 11:37:52 --- The Problem still persists with Wine version 1.2. This also effects the game Alien Swarm (Steam release).
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #74 from Henri Verbeet hverbeet@gmail.com 2010-07-21 07:38:39 --- Created an attachment (id=29733) --> (http://bugs.winehq.org/attachment.cgi?id=29733) patch
Please give the attached patch a try. It's of course not going to be as fast as disabling sRGB completely, but it should reduce the cost of RGB <-> sRGB transfers a bit.
http://bugs.winehq.org/show_bug.cgi?id=12453
Patrick Saalfeld suntrix.patrick@gmx.de changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |suntrix.patrick@gmx.de
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #75 from Ken Phillis Jr kphillisjr@gmail.com 2010-07-22 13:13:36 --- The patch does remove a lot of the conversion calls, however, the overall frame rate is still very low. With Left 4 Dead I'm seeing 25fps (even with the patch) instead of the expected 50 fps, however I do want to make sure it's noted that most of the sudden frame stutters have stopped especially when a large number of enemies are on screen. However, this could be accounted for by the fact i have most of the settings at max.
System Spec: Athlon II x4 630 Radeon hd 5770 Ubuntu 10.04 (64-bit) 4gb ddr3 memory Catalyst 10.6 drivers
http://bugs.winehq.org/show_bug.cgi?id=12453
V.platanias azathothgr@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |azathothgr@gmail.com
--- Comment #76 from V.platanias azathothgr@gmail.com 2010-07-23 05:27:41 --- The no-srgb patch blacks out most textures in Alien Swarm , but only after joining an online game for some reason. Before that, offline practice works fine. FPS goes from ~20-25 to ~50+ in lowest settings, 800x600.
The second patch makes most text disappear once I enter an online game, after which it stays invisible everywhere, with only random characters visible. FPS goes to about 30 to 40, again for offline practice. I tried spectating but I could no longer see the fps text. It seemed at about the same fps as offline.
Pentium dual-core 2.6GHz nvidia 8600 GT, driver v. 256.35 wine 1.2 ubuntu karmic, no compiz running
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #77 from beitler+winebugs@gmail.com 2010-07-23 07:53:30 --- Created an attachment (id=29787) --> (http://bugs.winehq.org/attachment.cgi?id=29787) Alien Swarm screenshot that shows the random text issue
wine 1.2-git + Henri's patch ubuntu 10.04, 64bit GeForce 8800 GT
http://bugs.winehq.org/show_bug.cgi?id=12453
eurekafag@eureka7.ru changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |eurekafag@eureka7.ru
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #78 from eurekafag@eureka7.ru 2010-07-24 10:20:13 --- Interesting thing: I've played online with second patch installed and had fonts visible and high FPS! That's because I hosted the game. Looks like the bug with lost symbols affects clients only but this still needs investigation.
http://bugs.winehq.org/show_bug.cgi?id=12453
mikey abc.mikey@googlemail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |abc.mikey@googlemail.com
--- Comment #79 from mikey abc.mikey@googlemail.com 2010-07-24 14:46:25 --- Hi there,
I've just had a chance to look at the last patch (attachment number 29733) and I'm getting a good speed boost on source games. L4D is able to run at 45-76 fps now instead of 20-37fps with minimal settings. And Portal is up to 35-75fps from 15-40fps on minimal settings.
However there are still times of considerable drops in speed (generally this is to the lower figure quoted).
http://bugs.winehq.org/show_bug.cgi?id=12453
Michael B toxatec@web.de changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |toxatec@web.de
--- Comment #80 from Michael B toxatec@web.de 2010-07-24 16:48:29 --- I have the same problems and improvements in Alien Swarm with sRGB disabled (older patch for wine-1.1.33), very similar to left4dead.
In both games I can get rid of the faults in multiplayer by changing video options, at least changing the "Shader detail" - I don't know if other options work. left4dead and Alien Swarm freeze with an ugly sound loop for some seconds once you apply the settings, but then everything looks OK and runs with nice framerates. As mentioned the menu strings sometimes disappear partly or completely, so it's not that easy to find the right options :/ But it works then.
Core2Duo 2.85GHz nVidia GeForce 8800GT (nvidia-drivers-256.35) wine-1.2 with Luke's 1.1.33 patch
http://bugs.winehq.org/show_bug.cgi?id=12453
Viktor Varland viktor@varland.se changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |viktor@varland.se
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #81 from Henri Verbeet hverbeet@gmail.com 2010-07-26 14:15:42 --- Patch should be in current git as 943fb2fb578b9d3d2fcd9c4e72bf465fde8e245d. I think the missing text in Alien Swarm etc. should no longer happen, but please test it a bit, it's a fairly invasive change.
http://bugs.winehq.org/show_bug.cgi?id=12453
byteframe byteframe@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |byteframe@gmail.com
--- Comment #82 from byteframe byteframe@gmail.com 2010-07-26 21:10:47 --- /me thought he was on cc list
Anyways, the latest git improves performance in source games' dxlevel >= 90 substantially. Without a dx80 fallback, I had to put Effect/Shader detail to their lowest values to get playable left 4 dead. I have left 4 dead running at max settings now (no aa), quite likely at with a faster framerate. Much appreciated!
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #83 from DL dredgingthelake@gmail.com 2010-07-27 02:59:09 --- I'm getting a massive performance increase with current git and Half-Life 2: Episode Two. On a particular save, the fps went from 14 fps to 66fps. Also, the disappearing HUD/menu problems appear to have been reduced (although they still occur occasionally).
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #84 from Henri Verbeet hverbeet@gmail.com 2010-07-27 03:46:03 --- (In reply to comment #83)
Episode Two. On a particular save, the fps went from 14 fps to 66fps. Also, the disappearing HUD/menu problems appear to have been reduced (although they still occur occasionally).
Is that a regression, or did it happen before? If it's a regression please do a regression test.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #85 from DL dredgingthelake@gmail.com 2010-07-27 03:48:42 ---
(In reply to comment #84)
Is that a regression, or did it happen before? If it's a regression please do a regression test.
The hud issue has been around for some time, but I have no idea how long. I'll try some old versions, and if the issue was not present, I'll run a regression test.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #86 from eurekafag@eureka7.ru 2010-07-27 09:50:49 --- As far as I see Alien Swarm still suffers text loss, nothing really changed comparing to the second patch. Didn't run actual multiplayer game but in lobby there is almost no text visible. When you're running a server everything shows well and runs fast (in offline mode and in multiplayer, too). The problem is only playing as client.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #87 from Fab netfab@gmail.com 2010-07-28 02:28:30 --- Created an attachment (id=29878) --> (http://bugs.winehq.org/attachment.cgi?id=29878) output from steam / Half-Life 2
Running Half-Life 2 with latest git seems to change nothing for me with fps. Moreover, I found lot of GL_INVALID_OPERATION in output log (see attached file). This does not happen with wine-1.2. I launch the game (simply) with :
$ wine Steam.exe -applaunch 220 -fullscreen -w 1024 -h 768 -dxlevel 90
nvidia-drivers 195.36.24
http://bugs.winehq.org/show_bug.cgi?id=12453
Fab netfab@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |netfab@gmail.com
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #88 from DL dredgingthelake@gmail.com 2010-07-28 02:34:42 --- (In reply to comment #87)
Created an attachment (id=29878)
--> (http://bugs.winehq.org/attachment.cgi?id=29878) [details]
output from steam / Half-Life 2
Running Half-Life 2 with latest git seems to change nothing for me with fps. Moreover, I found lot of GL_INVALID_OPERATION in output log (see attached file). This does not happen with wine-1.2. I launch the game (simply) with :
$ wine Steam.exe -applaunch 220 -fullscreen -w 1024 -h 768 -dxlevel 90
nvidia-drivers 195.36.24
The original Half-Life 2 (not the episodes) is not affected by this bug, as it uses an older version of the source engine. I've got 80+ average fps with Half-Life 2 for a long time. Of course, there are still the usual 25-50% performance hits that most games run under wine have, but this is not specific to Half-Life 2.
http://bugs.winehq.org/show_bug.cgi?id=12453
Alan A. grsfdhj@tiscali.it changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |grsfdhj@tiscali.it
--- Comment #89 from Alan A. grsfdhj@tiscali.it 2010-07-28 16:50:29 --- Tested Alien Swarm with latest GIT (just after having filed a test result using wine 1.2 stable :) ) and it works great. The game is now fluid, even putting every graphic option to the max. I noticed, though, that the "Multi core rendering" graphic option is not available anymore. Anyone else noticed this too?
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #90 from Alan A. grsfdhj@tiscali.it 2010-07-29 16:27:17 --- (In reply to comment #89)
Tested Alien Swarm with latest GIT (just after having filed a test result using wine 1.2 stable :) ) and it works great. The game is now fluid, even putting every graphic option to the max. I noticed, though, that the "Multi core rendering" graphic option is not available anymore. Anyone else noticed this too?
Had to get back to wine-1.2 stable, because the game was crashing often using wine-git. The "Multi core rendering" option reappeared but, surprisingly, setting it to disabled greatly boosted the performance. Now I can even turn effects to high with no slowdown.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #91 from byteframe byteframe@gmail.com 2010-07-31 15:48:30 --- I've noticed a problem with left4dead under wine-1.3.0, I don't know if it's strictly related the srgb fix. The high-quality flashlight effect (present when you have "Effects Quality" set to >= Medium) will not work in multiplayer, unless you turn effects quality down to low, and then return it back to Medium/High -- while remaining in a game. If one were to start a single player campaign the high quality flash light effect will work immediately.
http://bugs.winehq.org/show_bug.cgi?id=12453
Raymond superquad.vortex2@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |superquad.vortex2@gmail.com
--- Comment #92 from Raymond superquad.vortex2@gmail.com 2010-07-31 21:26:19 --- (In reply to comment #63)
Another performance factor in Source games is sound, it seems. Without any sound plugin, TF2 runs well on dxlevel 81 all high at ~15~35 fps (still too slow to be playable), but with the same settings and with the ALSA audio driver, I get 8~15 fps. I also get 10~20 fps with winepulse.
did you set your ALSA "default" device to dmix or hw instead of pulse when you are also using winepulse ?
http://bugs.winehq.org/show_bug.cgi?id=12453
Adam Bolte boltronics@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |boltronics@gmail.com
http://bugs.winehq.org/show_bug.cgi?id=12453
quaker quaker66@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |quaker66@gmail.com
--- Comment #93 from quaker quaker66@gmail.com 2010-08-08 12:03:05 --- i can confirm this to be fixed, performance is now perfect on my computer
Athlon II X2 240 2.8GHz 2048MB of RAM GeForce GTS250 1GB
my wineversion is latest git (shows as 1.3.0).
everything is perfectly smooth, in pre-1.2 versions it wasn't.
http://bugs.winehq.org/show_bug.cgi?id=12453
sh shooter0106@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |shooter0106@gmail.com
--- Comment #94 from sh shooter0106@gmail.com 2010-08-09 00:12:09 --- (In reply to comment #93)
i can confirm this to be fixed, performance is now perfect on my computer
Athlon II X2 240 2.8GHz 2048MB of RAM GeForce GTS250 1GB
my wineversion is latest git (shows as 1.3.0).
everything is perfectly smooth, in pre-1.2 versions it wasn't.
What games you tested? What is DirectX level you set?
http://bugs.winehq.org/show_bug.cgi?id=12453
Jaime Rave jaimerave@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC|jaimerave@gmail.com |
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #95 from quaker quaker66@gmail.com 2010-08-09 03:23:26 --- (In reply to comment #94)
(In reply to comment #93)
i can confirm this to be fixed, performance is now perfect on my computer
Athlon II X2 240 2.8GHz 2048MB of RAM GeForce GTS250 1GB
my wineversion is latest git (shows as 1.3.0).
everything is perfectly smooth, in pre-1.2 versions it wasn't.
What games you tested? What is DirectX level you set?
i tested Alien Swarm and Portal: First Slice, everything completely smooth as it should be, i didn't change dxlevel anyhow, so it uses 9 (in game settings, it shows both hardware and software directx level as 9). I'll test further in at least Left 4 Dead 2 and Half Life 2 Episode Two, and report back.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #96 from quaker quaker66@gmail.com 2010-08-09 03:25:08 --- (In reply to comment #94)
(In reply to comment #93)
i can confirm this to be fixed, performance is now perfect on my computer
Athlon II X2 240 2.8GHz 2048MB of RAM GeForce GTS250 1GB
my wineversion is latest git (shows as 1.3.0).
everything is perfectly smooth, in pre-1.2 versions it wasn't.
What games you tested? What is DirectX level you set?
btw, game details are all maxed, so i don't play at low or anything, on low it worked smoothly even before.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #97 from Ken Phillis Jr kphillisjr@gmail.com 2010-08-09 15:21:17 --- I can confirm that with wine 1.3.0 that the performance improved, however, left 4 dead still has missing text on screen, and this missing text can be bad enough to prevent one from knowing what the message says. As for directx level, i should be running at dx9 because that is the default for left 4 dead.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #98 from quaker quaker66@gmail.com 2010-08-09 15:23:45 --- (In reply to comment #97)
I can confirm that with wine 1.3.0 that the performance improved, however, left 4 dead still has missing text on screen, and this missing text can be bad enough to prevent one from knowing what the message says. As for directx level, i should be running at dx9 because that is the default for left 4 dead.
the performance is perfect for me :) i additionally tested also Fortress Forever mod, Half Life 2: Lost Coast and Half Life 2: Deathmatch; i can't test others atm. As for missing text, i can't say - I don't have Left 4 Dead atm. But that should belong to different bug, right? :)
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #99 from Henri Verbeet hverbeet@gmail.com 2010-08-09 15:51:16 --- (In reply to comment #98)
But that should belong to different bug, right? :)
Yeah, please try the patch attached to bug 23842 for that. Note that for AMD users there's also a Catalyst bug/regression involved with similar symptoms.
http://bugs.winehq.org/show_bug.cgi?id=12453
Artem S. Tashkinov t.artem@mailcity.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC|t.artem@mailcity.com |
http://bugs.winehq.org/show_bug.cgi?id=12453
jonathan-vola@hotmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |jonathan-vola@hotmail.com
--- Comment #100 from jonathan-vola@hotmail.com 2011-02-16 12:55:50 CST --- Still getting sRGB error with half life 2, wine 1.3.4, ubuntu lucid, and a geforce 9500 GT
fixme:d3d_surface:surface_load_location Downloading RGB surface 0x104d9e08 to reload it as sRGB.
The above message appears very often, but the main problem occurs rarely, for example, in route kanal.
When the train comes close enough (To load higher detail textures I presume) the game slows to a crawl (0.5fps or worse) and after several seconds the game returns to normal and the above output is spammed to the terminal.
The game once more slows when the train despawns at the other end of the tunnel.
Is there an easy way to stop this BS?
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #101 from Tobias Jakobi liquid.acid@gmx.net 2011-02-16 18:35:54 CST --- I don't know if this was posted here already, but this (upcoming?) GL extension could solve the sRGB dilemma: http://www.opengl.org/registry/specs/EXT/texture_sRGB_decode.txt
Some mesa drivers seem to implement this already, however AFAIK wine doesn't to use it yet.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #102 from Henri Verbeet hverbeet@gmail.com 2011-02-18 05:12:49 CST --- (In reply to comment #101)
I don't know if this was posted here already, but this (upcoming?) GL extension could solve the sRGB dilemma: http://www.opengl.org/registry/specs/EXT/texture_sRGB_decode.txt
Yes, that's what that extension is meant for. May take a while as I'll probably have to do it on my own time.
http://bugs.winehq.org/show_bug.cgi?id=12453
Nicky nheart@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |nheart@gmail.com
--- Comment #103 from Nicky nheart@gmail.com 2011-03-03 03:28:40 CST --- Any idea which nvidia GPUs support that extension (GL_EXT_texture_sRGB_decode): My nvidia ION (9400M) returns: [dheart@lightbringer ~]$ glxinfo | grep sRGB GLX_NV_float_buffer, GLX_ARB_fbconfig_float, GLX_EXT_framebuffer_sRGB GLX_EXT_texture_from_pixmap, GLX_EXT_framebuffer_sRGB, GLX_NV_float_buffer, GLX_ARB_fbconfig_float, GLX_EXT_framebuffer_sRGB, GL_ARB_framebuffer_object, GL_ARB_framebuffer_sRGB, GL_EXT_framebuffer_sRGB, GL_EXT_geometry_shader4, GL_EXT_texture_shared_exponent, GL_EXT_texture_sRGB,
no such extension. Any other nvidia users? Do you have that extension? Googling the matter doesn't really return usefull stuff. pp. Thanks for the patch, Henri, hope it gets accepted.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #104 from Henri Verbeet hverbeet@gmail.com 2011-03-03 04:35:11 CST --- (In reply to comment #103)
Any idea which nvidia GPUs support that extension (GL_EXT_texture_sRGB_decode): My nvidia ION (9400M) returns: [dheart@lightbringer ~]$ glxinfo | grep sRGB GLX_NV_float_buffer, GLX_ARB_fbconfig_float, GLX_EXT_framebuffer_sRGB GLX_EXT_texture_from_pixmap, GLX_EXT_framebuffer_sRGB, GLX_NV_float_buffer, GLX_ARB_fbconfig_float, GLX_EXT_framebuffer_sRGB, GL_ARB_framebuffer_object, GL_ARB_framebuffer_sRGB, GL_EXT_framebuffer_sRGB, GL_EXT_geometry_shader4, GL_EXT_texture_shared_exponent, GL_EXT_texture_sRGB,
no such extension. Any other nvidia users? Do you have that extension? Googling the matter doesn't really return usefull stuff.
It should start showing up in future nvidia driver versions, but git versions of Mesa are the only drivers I know of that currently support this.
http://bugs.winehq.org/show_bug.cgi?id=12453
lubosz lubosz@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |lubosz@gmail.com
--- Comment #105 from lubosz lubosz@gmail.com 2011-04-19 11:23:54 CDT --- bmonkey@bstation:~$ glxinfo | grep sRGB GLX_NV_float_buffer, GLX_ARB_fbconfig_float, GLX_EXT_framebuffer_sRGB GLX_EXT_texture_from_pixmap, GLX_EXT_framebuffer_sRGB, GLX_NV_float_buffer, GLX_ARB_fbconfig_float, GLX_EXT_framebuffer_sRGB, GL_ARB_framebuffer_object, GL_ARB_framebuffer_sRGB, GL_EXT_framebuffer_sRGB, GL_EXT_geometry_shader4, GL_EXT_texture_shared_exponent, GL_EXT_texture_sRGB, bmonkey@bstation:~$ nvidia-settings --version
nvidia-settings: version 270.41.03 (buildmeister@swio-display-x86-rhel47-07.nvidia.com) Sat Apr 9 00:26:16 PDT 2011 The NVIDIA X Server Settings tool.
This program is used to configure the NVIDIA Linux graphics driver. For more detail, please see the nvidia-settings(1) man page.
Copyright (C) 2004 - 2010 NVIDIA Corporation.
http://bugs.winehq.org/show_bug.cgi?id=12453
Alexey Loukianov mooroon2@mail.ru changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |mooroon2@mail.ru
--- Comment #106 from Alexey Loukianov mooroon2@mail.ru 2011-09-11 17:14:02 CDT --- Here is a lazy reminder to keep this bug "active": texture redownload for sRGB<=>linear conversion still happen with orm=backbuffer as of Wine 1.3.28.
GL_EXT_texture_sRGB_decode driver support still sucks: http://feedback.wildfiregames.com/report/opengl/feature/GL_EXT_texture_sRGB_...
What's interesting is that nVIDIA officially mentions this extension in FXAA PDF brief when it comes to porting FXAA to the OpenGL from DX10/11. FXAA brief I'm talking about was published towards the end of 2010. Now we're approaching the end of 2011 and this extension is not supported by nVIDIA OpenGL ICD neither on Windows nor on Linux.
Besides I don't get noticeable FPS drop in HL2:EP1 caused by this re-downloads when playing with orm=backbuffer (it looks like that with orm=fbo no re-downloads happen). It might be caused by recent changes in Wine 1.3.27 made orm=backbuffer several times slower in general comparing to earlier versions. This general slowdown might "mask" the slowdown caused by sRGB redownloads/conversions.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #107 from Alexey Loukianov mooroon2@mail.ru 2011-10-18 16:32:12 CDT --- Installed nVIDIA Linux 32bit Driver v.285.05.09 and got this:
[lexa2@lx2linux ddd]$ glxinfo | grep sRGB GLX_NV_float_buffer, GLX_ARB_fbconfig_float, GLX_EXT_framebuffer_sRGB GLX_EXT_texture_from_pixmap, GLX_EXT_framebuffer_sRGB, GLX_NV_float_buffer, GLX_ARB_fbconfig_float, GLX_EXT_framebuffer_sRGB, GL_ARB_framebuffer_sRGB, GL_ARB_geometry_shader4, GL_EXT_framebuffer_sRGB, GL_EXT_geometry_shader4, GL_EXT_texture_shared_exponent, GL_EXT_texture_sRGB, GL_EXT_texture_sRGB_decode, GL_EXT_texture_storage,
Thus, GL_EXT_texture_sRGB_decode now seems to be supported, at least on the nVIDIA side.
Henri, had there been any code commited into Wine that makes use of this extension? If the fix for sRGB<=>RBG re-download is already in place I could run some tests in HL2:EP1.
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #108 from Alexey Loukianov mooroon2@mail.ru 2011-10-18 22:03:02 CDT --- PS/FYI on topic: I had spotted at least two regression with apps running under Wine when using nVIDIA 825.05.09 driver. Here is report I posted to official nVIDIA Linux support forum: http://www.nvnews.net/vbulletin/showthread.php?p=2492073 .
Also I had asked there about support for GLX_EXT_swap_control_tear and was told that they plan to add it soon: http://www.nvnews.net/vbulletin/showthread.php?t=167470
http://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #109 from Henri Verbeet hverbeet@gmail.com 2011-10-19 08:00:04 CDT --- (In reply to comment #107)
Henri, had there been any code commited into Wine that makes use of this extension?
75c8e9f
http://bugs.winehq.org/show_bug.cgi?id=12453
giuseppegebbia@msn.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |giuseppegebbia@msn.com
http://bugs.winehq.org/show_bug.cgi?id=12453
salamander purake salamanderrake@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |salamanderrake@gmail.com
--- Comment #110 from salamander purake salamanderrake@gmail.com 2012-11-21 18:52:30 CST --- How can this be reopened if there is no reason given. Also shouldn't the relevent patches be added to git by now.
http://bugs.winehq.org/show_bug.cgi?id=12453
butraxz@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |butraxz@gmail.com
--- Comment #111 from butraxz@gmail.com 2013-04-22 14:52:54 CDT --- Is this still an issue in current(1.5.28) or newer wine ? You may also close this as abandoned if you feel that that this is issue is no longer relevant to you.
http://bugs.winehq.org/show_bug.cgi?id=12453
Jarkko K jarkko_korpi@hotmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |jarkko_korpi@hotmail.com
--- Comment #112 from Jarkko K jarkko_korpi@hotmail.com --- Ati 7870, 14.3beta driver
glxinfo | grep sRGB GLX_EXT_framebuffer_sRGB, GLX_EXT_import_context, GLX_EXT_swap_control, GL_ARB_framebuffer_sRGB, GL_ARB_geometry_shader4, GL_EXT_framebuffer_sRGB, GL_EXT_geometry_shader4, GL_EXT_texture_object, GL_EXT_texture_rectangle, GL_EXT_texture_sRGB, GL_EXT_texture_sRGB_decode, GL_EXT_texture_shared_exponent, GL_ARB_framebuffer_sRGB, GL_ARB_geometry_shader4, GL_EXT_framebuffer_sRGB, GL_EXT_geometry_shader4, GL_EXT_texture_object, GL_EXT_texture_rectangle, GL_EXT_texture_sRGB, GL_EXT_texture_sRGB_decode, GL_EXT_texture_shared_exponent,
Is that sRGB_decode used now?
https://bugs.winehq.org/show_bug.cgi?id=12453
joaopa jeremielapuree@yahoo.fr changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |jeremielapuree@yahoo.fr
--- Comment #113 from joaopa jeremielapuree@yahoo.fr --- Still an issue with current wine?
https://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #114 from joaopa jeremielapuree@yahoo.fr --- This bug can be closed as ABANDONED, I think....
https://bugs.winehq.org/show_bug.cgi?id=12453
winetest@luukku.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |00cpxxx@gmail.com, | |austinenglish@gmail.com, | |winetest@luukku.com
--- Comment #115 from winetest@luukku.com --- (In reply to salamander purake from comment #110)
How can this be reopened if there is no reason given. Also shouldn't the relevent patches be added to git by now.
(In reply to joaopa from comment #114)
This bug can be closed as ABANDONED, I think....
Should be closed definitely...see comment history.
https://bugs.winehq.org/show_bug.cgi?id=12453
winetest@luukku.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |dimesio@earthlink.net
https://bugs.winehq.org/show_bug.cgi?id=12453
--- Comment #116 from joaopa jeremielapuree@yahoo.fr --- Does the bug still occur with wine4.0-rc7?
https://bugs.winehq.org/show_bug.cgi?id=12453
Matteo Bruni matteo.mystral@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Resolution|--- |FIXED Fixed by SHA1| |75c8e9f7b42173eb3c08e4ae87d | |485caca431a9b Status|REOPENED |RESOLVED
--- Comment #117 from Matteo Bruni matteo.mystral@gmail.com --- Let's close this bug, the original / main issue was repeated sRGB conversions and that's long fixed.
https://bugs.winehq.org/show_bug.cgi?id=12453
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #118 from Alexandre Julliard julliard@winehq.org --- Closing bugs fixed in 4.2.
-
wine-bugs＠winehq.org