Re: [r200] Mesa CVS text and texture bug
Dieter NÃtzel wrote: I get this even with X.org CVS. The texture problem is not persistent but the text 'of by oen or two'. http://www.nuetzel-hh.de/public/Celestia-Mesa-CSV-r200.png What version of Celestia is this? I get neither of these bugs here (celestia 1.3.0). Roland --- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_ide95&alloc_id396&op=click -- ___ Dri-devel mailing list Dri-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [r200] Mesa CVS text and texture bug
Am Freitag, 11. Februar 2005 23:12 schrieb Roland Scheidegger: > Dieter NÃtzel wrote: > > I get this even with X.org CVS. > > The texture problem is not persistent but the text 'of by oen or two'. > > > > http://www.nuetzel-hh.de/public/Celestia-Mesa-CSV-r200.png > > What version of Celestia is this? I get neither of these bugs here > (celestia 1.3.0). 1.3.2 (latest release from August 2004, self compiled) Mesa (LIBGL_ALWAYS_INDIRECT) works OK. BTW There is _BIG_ regression with quake3 (quake3-smp)! quake3-smp went down from ~198 - ~142 fps (with and without HIGH textures) !!! Most worse with SMP. !!! Do you have any texture load changes committed on Thursday to Friday? -Dieter --- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_ide95&alloc_id396&op=click -- ___ Dri-devel mailing list Dri-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [r200] Mesa CVS text and texture bug
Am Montag, 14. Februar 2005 14:37 schrieb Dieter NÃtzel: > Am Freitag, 11. Februar 2005 23:12 schrieb Roland Scheidegger: > > Dieter NÃtzel wrote: > > > I get this even with X.org CVS. > > > The texture problem is not persistent but the text 'of by oen or two'. > BTW There is _BIG_ regression with quake3 (quake3-smp)! > > quake3-smp went down from ~198 - ~142 fps (with and without HIGH textures) > > !!! Most worse with SMP. !!! > > Do you have any texture load changes committed on Thursday to Friday? This on is SOLVED! It was due to SuSE's CKRM_CPU_SCHEDULE scheduler. Back to Linux's 'normal' one solve it. quake3-smp ~198 fps 2. high ~193 fps HIGH 640x480 window on 1280x1024x24/32. 1024x768 window give ~100 fps, again. -Dieter --- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_ide95&alloc_id396&op=click -- ___ Dri-devel mailing list Dri-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [r200] Mesa CVS text and texture bug
Dieter NÃtzel wrote: This on is SOLVED! It was due to SuSE's CKRM_CPU_SCHEDULE scheduler. Back to Linux's 'normal' one solve it. quake3-smp ~198 fps 2. high ~193 fps HIGH 640x480 window on 1280x1024x24/32. 1024x768 window give ~100 fps, again. This is interesting, though. The use of a different scheduler should probably not have such a huge impact on performance (if no other cpu or io-heavy processes are running). Are you using pageflip? If no the r200CopyBuffer calls sched_yield() which potentially could cause such a difference. But if you're using pageflip I'm not sure why there would be such a drastic difference. Roland --- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_ide95&alloc_id396&op=click -- ___ Dri-devel mailing list Dri-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [r200] Mesa CVS text and texture bug
Am Montag, 14. Februar 2005 19:40 schrieb Roland Scheidegger: > Dieter NÃtzel wrote: > > This on is SOLVED! > > > > It was due to SuSE's CKRM_CPU_SCHEDULE scheduler. > > > > Back to Linux's 'normal' one solve it. > > > > quake3-smp > > ~198 fps 2. high > > ~193 fps HIGH > > > > 640x480 window on 1280x1024x24/32. > > > > 1024x768 window give ~100 fps, again. > > This is interesting, though. The use of a different scheduler should > probably not have such a huge impact on performance (if no other cpu or > io-heavy processes are running). > Are you using pageflip? All time long. > If no the r200CopyBuffer calls sched_yield() > which potentially could cause such a difference. But if you're using > pageflip I'm not sure why there would be such a drastic difference. What next? ;-) What about AGP 1x, 2x, 4x, etc.? I have "auto" in my BIOS settings. Section "Device" BoardName"Radeon 8500 QL" BusID"1:5:0" Driver "radeon" Identifier "Device[0]" Option "EnablePageflip" Option "DPMS" Option"AGPFastWrite" "1" Option"AGPMode" "4" Screen 0 Option "Rotate" "on" VendorName "ATI" EndSection dmesg show this: agpgart: Found an AGP 2.0 compliant device at :00:00.0. agpgart: Putting AGP V2 device at :00:00.0 into 2x mode agpgart: Putting AGP V2 device at :01:05.0 into 2x mode glxinfo OpenGL vendor string: Tungsten Graphics, Inc. OpenGL renderer string: Mesa DRI R200 20041207 AGP 4x x86/MMX+/3DNow!+/SSE TCL OpenGL version string: 1.3 Mesa 6.3 So only AGPGART output is wrong? How have you tested "local" (texdown)? Best here is: 64 x 256 Border 0 (1 give sig fault, but was OK several months ago) Format: GL_RGB Type: GL_UNSIGNED_SHORT_5_6_6 IntFormat: GL_RGB Bias: No TexSubImage: Yes (!) => 373.289 MB/second Border _doubled_ it in the past. -Dieter --- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_ide95&alloc_id396&op=click -- ___ Dri-devel mailing list Dri-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [r200] Mesa CVS text and texture bug
Am Montag, 14. Februar 2005 20:06 schrieb Dieter NÃtzel: > Am Montag, 14. Februar 2005 19:40 schrieb Roland Scheidegger: > > Dieter NÃtzel wrote: > > > This on is SOLVED! > > > > > > It was due to SuSE's CKRM_CPU_SCHEDULE scheduler. > > > > > > Back to Linux's 'normal' one solve it. > > > > > > quake3-smp > > > ~198 fps 2. high > > > ~193 fps HIGH > > > > > > 640x480 window on 1280x1024x24/32. > > > > > > 1024x768 window give ~100 fps, again. > > > > This is interesting, though. The use of a different scheduler should > > probably not have such a huge impact on performance (if no other cpu or > > io-heavy processes are running). > > Are you using pageflip? > > All time long. > > > If no the r200CopyBuffer calls sched_yield() > > which potentially could cause such a difference. But if you're using > > pageflip I'm not sure why there would be such a drastic difference. > > What next? ;-) > > What about AGP 1x, 2x, 4x, etc.? > > I have "auto" in my BIOS settings. > > Section "Device" > BoardName"Radeon 8500 QL" > BusID"1:5:0" > Driver "radeon" > Identifier "Device[0]" > Option "EnablePageflip" > Option "DPMS" > Option"AGPFastWrite" "1" > Option"AGPMode" "4" > Screen 0 > Option "Rotate" "on" > VendorName "ATI" > EndSection > > dmesg show this: > agpgart: Found an AGP 2.0 compliant device at :00:00.0. > agpgart: Putting AGP V2 device at :00:00.0 into 2x mode > agpgart: Putting AGP V2 device at :01:05.0 into 2x mode > > glxinfo > OpenGL vendor string: Tungsten Graphics, Inc. > OpenGL renderer string: Mesa DRI R200 20041207 AGP 4x x86/MMX+/3DNow!+/SSE > TCL OpenGL version string: 1.3 Mesa 6.3 > > So only AGPGART output is wrong? > > How have you tested "local" (texdown)? > > Best here is: > 64 x 256 Border 0 (1 give sig fault, but was OK several months ago) > Format: GL_RGB > Type: GL_UNSIGNED_SHORT_5_6_6 > IntFormat: GL_RGB > Bias: No > TexSubImage: Yes (!) > > => 373.289 MB/second > > Border _doubled_ it in the past. Reading symbols from /usr/X11R6/lib/modules/dri/r200_dri.so...done. Loaded symbols for /usr/X11R6/lib/modules/dri/r200_dri.so Reading symbols from /usr/X11R6/lib/libexpat.so.0...done. Loaded symbols for /usr/X11R6/lib/libexpat.so.0 Reading symbols from /usr/lib/libtxc_dxtn.so...done. Loaded symbols for /usr/lib/libtxc_dxtn.so Reading symbols from /usr/X11R6/lib/X11/locale/lib/common/xlcDef.so.2...done. Loaded symbols for /usr/X11R6/lib/X11/locale/lib/common/xlcDef.so.2 #0 0x405e9dc5 in _tnl_build_vertices () from /usr/X11R6/lib/modules/dri/r200_dri.so (gdb) bt #0 0x405e9dc5 in _tnl_build_vertices () from /usr/X11R6/lib/modules/dri/r200_dri.so #1 0x405e3513 in run_render () from /usr/X11R6/lib/modules/dri/r200_dri.so #2 0x405d36f2 in _tnl_run_pipeline () from /usr/X11R6/lib/modules/dri/r200_dri.so #3 0x40550b0f in r200WrapRunPipeline () from /usr/X11R6/lib/modules/dri/r200_dri.so #4 0x405f25cb in _tnl_flush_vtx () from /usr/X11R6/lib/modules/dri/r200_dri.so #5 0x405ed7e5 in _tnl_FlushVertices () from /usr/X11R6/lib/modules/dri/r200_dri.so #6 0x40563b73 in r200FlushVertices () from /usr/X11R6/lib/modules/dri/r200_dri.so #7 0x405b960f in _mesa_TexSubImage2D () from /usr/X11R6/lib/modules/dri/r200_dri.so #8 0x0804932c in MeasureDownloadRate () at texdown.c:158 #9 0x08049675 in Display () at texdown.c:256 #10 0x40045f8e in processWindowWorkList () from /usr/lib/libglut.so.3 #11 0x4004601a in __glutProcessWindowWorkLists () from /usr/lib/libglut.so.3 #12 0x4004608b in glutMainLoop () from /usr/lib/libglut.so.3 #13 0x08049b21 in main (argc=1, argv=0xbfffee44) at texdown.c:384 -Dieter --- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_ide95&alloc_id396&op=click -- ___ Dri-devel mailing list Dri-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [r200] Mesa CVS text and texture bug
Am Montag, 14. Februar 2005 19:40 schrieb Roland Scheidegger: > Dieter NÃtzel wrote: > > This on is SOLVED! > > > > It was due to SuSE's CKRM_CPU_SCHEDULE scheduler. > > > > Back to Linux's 'normal' one solve it. > > > > quake3-smp > > ~198 fps 2. high > > ~193 fps HIGH > > > > 640x480 window on 1280x1024x24/32. > > > > 1024x768 window give ~100 fps, again. > > This is interesting, though. The use of a different scheduler should > probably not have such a huge impact on performance (if no other cpu or > io-heavy processes are running). > Are you using pageflip? If no the r200CopyBuffer calls sched_yield() > which potentially could cause such a difference. But if you're using > pageflip I'm not sure why there would be such a drastic difference. Try "latencytest0.42-png". Especially have a look at "top" scheduler numbers. -Dieter --- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_ide95&alloc_id396&op=click -- ___ Dri-devel mailing list Dri-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [r200] Mesa CVS text and texture bug
Am Freitag, 11. Februar 2005 22:55 schrieb Dieter NÃtzel: > I get this even with X.org CVS. > The texture problem is not persistent but the text 'of by oen or two'. > > http://www.nuetzel-hh.de/public/Celestia-Mesa-CSV-r200.png Texture "bug" is due to texture units. It arise with "texture_units" > 4. => to few texture memory? Text corruption still there. -Dieter --- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_ide95&alloc_id396&op=click -- ___ Dri-devel mailing list Dri-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [r200] Mesa CVS text and texture bug
On Mon, Feb 14, 2005 at 07:40:20PM +0100, Roland Scheidegger wrote: > Dieter Nützel wrote: > >Back to Linux's 'normal' one solve it. > > This is interesting, though. The use of a different scheduler should > probably not have such a huge impact on performance (if no other cpu or > io-heavy processes are running). > Are you using pageflip? If no the r200CopyBuffer calls sched_yield() > which potentially could cause such a difference. But if you're using > pageflip I'm not sure why there would be such a drastic difference. Isn't sched_yield() a really bad idea under 2.6 kernels if you actually wan't a lot of CPU time overall? It puts the process on the back of the expired queue, which means IIRC that it won't get another timeslice until every process in the active queue has expired. cheers, Phil -- http://www.kantaka.co.uk/ .oOo. public key: http://www.kantaka.co.uk/gpg.txt --- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click -- ___ Dri-devel mailing list Dri-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [r200] Mesa CVS text and texture bug
On Mon, 2005-02-14 at 20:06 +0100, Dieter NÃtzel wrote: > > dmesg show this: > agpgart: Found an AGP 2.0 compliant device at :00:00.0. > agpgart: Putting AGP V2 device at :00:00.0 into 2x mode > agpgart: Putting AGP V2 device at :01:05.0 into 2x mode > > glxinfo > OpenGL vendor string: Tungsten Graphics, Inc. > OpenGL renderer string: Mesa DRI R200 20041207 AGP 4x x86/MMX+/3DNow!+/SSE TCL > OpenGL version string: 1.3 Mesa 6.3 > > So only AGPGART output is wrong? Rather the other way around. Usually, this means either the AGP bridge or the card don't actually support more than 2x. -- Earthling Michel DÃnzer | Debian (powerpc), X and DRI developer Libre software enthusiast| http://svcs.affero.net/rm.php?r=daenzer --- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_ide95&alloc_id396&op=click -- ___ Dri-devel mailing list Dri-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [r200] Mesa CVS text and texture bug
Am Montag, 14. Februar 2005 21:49 schrieb Michel DÃnzer: > On Mon, 2005-02-14 at 20:06 +0100, Dieter NÃtzel wrote: > > dmesg show this: > > agpgart: Found an AGP 2.0 compliant device at :00:00.0. > > agpgart: Putting AGP V2 device at :00:00.0 into 2x mode > > agpgart: Putting AGP V2 device at :01:05.0 into 2x mode > > > > glxinfo > > OpenGL vendor string: Tungsten Graphics, Inc. > > OpenGL renderer string: Mesa DRI R200 20041207 AGP 4x > > x86/MMX+/3DNow!+/SSE TCL OpenGL version string: 1.3 Mesa 6.3 > > > > So only AGPGART output is wrong? > > Rather the other way around. Usually, this means either the AGP bridge > or the card don't actually support more than 2x. You are right. AMD 768MPX support AGP Pro. => max 4x My r200 => max 2x (?) But lspci -vv looks wrong. See the AMD docs. http://www.msi-computer.de/produkte/main_idx_view.php?Prod_id=191 I see 4x max in the BIOS. 00:00.0 Host bridge: Advanced Micro Devices [AMD] AMD-760 MP [IGD4-2P] System Controller (rev 11) Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- Status: Cap+ 66Mhz+ UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- SERR- TAbort- SERR- Reset- FastB2B- 01:05.0 VGA compatible controller: ATI Technologies Inc Radeon R200 QL [Radeon 8500 LE] (prog-if 00 [VGA]) Subsystem: C.P. Technology Co. Ltd: Unknown device 2034 Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping+ SERR- FastB2B- Status: Cap+ 66Mhz+ UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- SERR- [disabled] [size=128K] Capabilities: [58] AGP version 2.0 Status: RQ=48 Iso- ArqSz=0 Cal=0 SBA+ ITACoh- GART64- HTrans- 64bit- FW+ AGP3- Rate=x1,x2 Command: RQ=16 ArqSz=0 Cal=0 SBA+ AGP+ GART64- 64bit- FW- Rate=x2 Capabilities: [50] Power Management version 2 Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-) Status: D0 PME-Enable- DSel=0 DScale=0 PME- Michel, what about the Celestia text bug. Seems to be appear only with 'small' text. http://www.nuetzel-hh.de/public/Celestia-Mesa-CSV-r200.png LIBGL_ALWAYS_INDIRECT is fine. Thanks, Dieter --- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_ide95&alloc_id396&op=click -- ___ Dri-devel mailing list Dri-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dri-devel
