|
|
Subject: Re: Performance problem with radeon 9200SE PCI -
msg#00069
Helge Hafting wrote:
One radeon 9200SE PCI, doing 1280x1024 24-bit. Tuxracer works, but is
too choppy at 1280x1024. Of course I can lower resolution, but I
hope for more. pinball gets 24fps, other games are also noticeably
slower glxgears gets 394fps, or 59fps maximized.
Cpu: AMD opteron 1.8GHz
Well, the 9200SE is a pretty slow card to begin with (should still not
be slower than a G550 though). However, there are known performance
problems with the pci based R200-RV280 cards with the dri drivers.
Unfortunately, no one knows why :-(.
I don't know much about configuring for 3D
performance. I have seen some tips about setting the chipID. X
identifies it as chipID 5964. There are recommendations for
ovverriding with 0x514D to get better 3D performance, it makes no
difference at all for me.
I don't think there's anything you could do, you might try newest
drivers from Mesa cvs. Not sure what version you're really using,
XFree86 4.3 and Mesa 6.1 typically aren't packaged together...
Changing chipids will do nothing and should be avoided, it will only
help if the driver doesn't detect the card correctly if it's too new.
Roland
-------------------------------------------------------
This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
Project Admins to receive an Apple iPod Mini FREE for your judgement on
who ports your project to Linux PPC the best. Sponsored by IBM.
Deadline: Sept. 24. Go here: http://sf.net/ppc_contest.php
--
Thread at a glance:
Previous Message by Date:
Re: S3 savage IX - segfault on modprobe
> I plan to commmit the merge of the DRI savage driver as sson as the
> xorg cvs tree is unlocked. I haven't gotten a chance to look into the
> mode problems you were seeing, but i have a few ideas. Besides that
> I've been working on a substantial rewrite of the DDX (fix Xv,
> dualhead, mergedfb, dvi, etc.), but that's still a work in progress.
>
> Alex
Well that sounds great, I've downloaded X.org cvs now and compiled and
installed that along with the drm savage module from dri cvs with pretty
much the same results as before. (compliants about agp in the
Xorg.0.log) I'm happy to test any code you produce as I have got access
to a fast enough computer to compile it all on now.
Keep up the good work
Glyn
-------------------------------------------------------
This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
Project Admins to receive an Apple iPod Mini FREE for your judgement on
who ports your project to Linux PPC the best. Sponsored by IBM.
Deadline: Sept. 24. Go here: http://sf.net/ppc_contest.php
--
Next Message by Date:
Re: Savage: unresolved symbols (drmCreateContext,drmAuthMagic,drmDestroyContext,drmRmMap,DRICreatePCIBusID)
Hi Devon,
your Xorg does work, doesn't it? You're getting unresolved symbols
because you don't have dri enabled. The errors are harmless, though I
think the driver could be made nicer, so that it doesn't trigger these
messages if DRI is disabled.
I don't understand why DRI is disabled though. You have the necessary
Xserver modules loaded (dri and glx). Could you send your full server
log? Maybe it gives some more hints.
The warnings about write combining are harmless too. Ignore them.
Regards,
Felix
On Sun, 19 Sep 2004 16:41:06 -0400
Devon Miller <devon.miller2@xxxxxxxxxxx> wrote:
> I'm running Gentoo and I unmerged opengl-update before starting.
> I've checked out the current head branch from CvS.
> It builds and installs without errors.
> During startup, I get these errors:
>
> <>
>
> > <>(II) SAVAGE: driver (version 1.1.27a) for S3 Savage chipsets: Savage4,
> > Savage3D, Savage3D-MV, Savage2000, Savage/MX-MV, Savage/MX,
> > Savage/IX-MV, Savage/IX, ProSavage PM133, ProSavage KM133,
> > ProSavage PN133, ProSavage KN133, SuperSavage/MX 128,
> > SuperSavage/MX 64, SuperSavage/MX 64C, SuperSavage/IX 128,
> > SuperSavage/IX 128, SuperSavage/IX 64, SuperSavage/IX 64,
> > SuperSavage/IXC 64, SuperSavage/IXC 64, ProSavage DDR,
> > ProSavage DDR-K
>
> [...]
>
> > (**) SAVAGE(0): Depth 24, (--) framebuffer bpp 32
> > (==) SAVAGE(0): RGB weight 888
> > (==) SAVAGE(0): Default visual is TrueColor
> > (II) SAVAGE(0): vgaHWGetIOBase: hwp->IOBase is 0x03d0, hwp->PIOOffset
> > is 0x0000
> > (==) SAVAGE(0): Using AGP 1x mode
> > (II) SAVAGE(0): Using 16 MB AGP aperture
> > (==) SAVAGE(0): Using HW cursor
> > (==) SAVAGE(0): Using video BIOS to set modes
>
> [...]
>
> > (II) SAVAGE(0): VESA BIOS detected
> > (II) SAVAGE(0): VESA VBE Version 2.0
> > (II) SAVAGE(0): VESA VBE Total Mem: 15296 kB
> > (II) SAVAGE(0): VESA VBE OEM: S3 Incorporated. Twister BIOS
> > (II) SAVAGE(0): VESA VBE OEM Software Rev: 1.0
> > (II) SAVAGE(0): VESA VBE OEM Vendor: S3 Incorporated.
> > (II) SAVAGE(0): VESA VBE OEM Product: VBE 2.0
> > (II) SAVAGE(0): VESA VBE OEM Product Rev: Rev 1.1
> > (--) SAVAGE(0): Chip: id 8d02, "ProSavage KN133"
> > (--) SAVAGE(0): Engine: "Twister"
> > (--) SAVAGE(0): mapping MMIO @ 0xe0000000 with size 0x80000
> > (==) SAVAGE(0): Using gamma correction (1.0, 1.0, 1.0)
> > (--) SAVAGE(0): probed videoram: 16384k
>
> [...]
>
> > (II) SAVAGE(0): initializing int10
> > (II) SAVAGE(0): Primary V_BIOS segment is: 0xc000
> > (II) SAVAGE(0): VESA BIOS detected
> > (II) SAVAGE(0): VESA VBE Version 2.0
> > (II) SAVAGE(0): VESA VBE Total Mem: 15296 kB
> > (II) SAVAGE(0): VESA VBE OEM: S3 Incorporated. Twister BIOS
> > (II) SAVAGE(0): VESA VBE OEM Software Rev: 1.0
> > (II) SAVAGE(0): VESA VBE OEM Vendor: S3 Incorporated.
> > (II) SAVAGE(0): VESA VBE OEM Product: VBE 2.0
> > (II) SAVAGE(0): VESA VBE OEM Product Rev: Rev 1.1
> > (--) SAVAGE(0): mapping framebuffer @ 0x90000000 with size 0x1000000
> > (==) SAVAGE(0): Write-combining range (0x90000000,0x1000000)
> > (II) SAVAGE(0): Splitting WC range: base: 0x92000000, size: 0x5000000
> > (II) SAVAGE(0): Splitting WC range: base: 0x94000000, size: 0x3000000
>
> This is marked as a warning. What does it mean?
>
> > (==) SAVAGE(0): Write-combining range (0x96000000,0x1000000)
> > (WW) SAVAGE(0): Failed to set up write-combining range
> > (0x94000000,0x3000000)
> > (WW) SAVAGE(0): Failed to set up write-combining range
> > (0x92000000,0x5000000)
> > (II) SAVAGE(0): map aperture:0x413db000
> > (--) SAVAGE(0): Chose mode 118 at 60Hz.
> > (II) SAVAGE(0): 9348 kB of Videoram needed for 3D; 16384 kB of
> > Videoram available
> > (II) SAVAGE(0): Sufficient Videoram available for 3D
> > (EE) SAVAGE(0): DRI isn't enabled
> > (II) SAVAGE(0): Using 1280 lines for offscreen memory.
> > (II) SAVAGE(0): Using XFree86 Acceleration Architecture (XAA)
> > Screen to screen bit blits
> > Solid filled rectangles
> > 8x8 mono pattern filled rectangles
> > Indirect CPU to Screen color expansion
> > Solid Lines
> > Image Writes
> > Offscreen Pixmaps
> > Setting up tile and stipple cache:
> > 32 128x128 slots
> > 12 256x256 slots
> > (==) SAVAGE(0): Backing store disabled
> > (**) Option "dpms" "true"
> > (**) SAVAGE(0): DPMS enabled
> > (EE) SAVAGE(0): Direct rendering disabled
> > (EE) SAVAGE(0): Direct rendering disabled
>
> I believe this to be the cause of DRM being disabled:
>
> > Symbol drmCreateContext from module
> > /usr/X11R6/lib/modules/drivers/savage_drv.o is unresolved!
> > Symbol drmAuthMagic from module
> > /usr/X11R6/lib/modules/drivers/savage_drv.o is unresolved!
> > Symbol drmDestroyContext from module
> > /usr/X11R6/lib/modules/drivers/savage_drv.o is unresolved!
> > Symbol drmRmMap from module
> > /usr/X11R6/lib/modules/drivers/savage_drv.o is unresolved!
> > Symbol drmRmMap from module
> > /usr/X11R6/lib/modules/drivers/savage_drv.o is unresolved!
> > Symbol drmRmMap from module
> > /usr/X11R6/lib/modules/drivers/savage_drv.o is unresolved!
> > Symbol drmRmMap from module
> > /usr/X11R6/lib/modules/drivers/savage_drv.o is unresolved!
> > Symbol DRICreatePCIBusID from module
> > /usr/X11R6/lib/modules/drivers/savage_drv.o is unresolved!
>
>
> All of these symbols are defined in xf86drm.c, but I have no idea how to
> figure out why they're not resolving.
> What must I do to make these resolve?
>
> Devon
>
> PS: This is my xorg.conf (sans comments for brevity)
>
> > Section "Module"
> > Load "dbe"
> > SubSection "extmod"
> > Option "omit xfree86-dga"
> > EndSubSection
> > Load "type1"
> > Load "speedo"
> > Load "freetype"
> > Load "xtt"
> > Load "synaptics"
> > Load "glx"
> > Load "dri"
> > EndSection
> > Section "Files"
> > RgbPath "/usr/X11R6/lib/X11/rgb"
> > FontPath "/usr/share/fonts/misc/"
> > FontPath "/usr/share/fonts/TTF/"
> > FontPath "/usr/share/fonts/Speedo/"
> > FontPath "/usr/share/fonts/Type1/"
> > FontPath "/usr/share/fonts/CID/"
> > FontPath "/usr/share/fonts/75dpi/"
> > FontPath "/usr/share/fonts/100dpi/"
> > FontPath "/usr/share/fonts/local/"
> > FontPath "/usr/share/fonts/TrueType/"
> > FontPath "/usr/share/fonts/freefont/"
> > EndSection
> > Section "ServerFlags"
> > EndSection
> > Section "InputDevice"
> > Identifier "Keyboard1"
> > Driver "Keyboard"
> > Option "AutoRepeat" "500 30"
> > Option "XkbRules" "xorg"
> > Option "XkbModel" "pc104"
> > Option "XkbLayout" "us"
> > EndSection
> > Section "InputDevice"
> > Identifier "Mouse1"
> > Driver "mouse"
> > Option "Protocol" "Auto"
> > Option "Device" "/dev/mouse"
> > Option "Emulate3Buttons"
> > EndSection
> > Section "InputDevice"
> > Identifier "Synaptic Mouse"
> > Driver "synaptics"
> > Option "Device" "/dev/psaux"
> > Option "Protocol" "auto-dev"
> > Option "LeftEdge" "1700"
> > Option "RightEdge" "5300"
> > Option "TopEdge" "1700"
> > Option "BottomEdge" "4200"
> > Option "FingerLow" "25"
> > Option "FingerHigh" "30"
> > Option "MaxTapTime" "180"
> > Option "MaxTapMove" "220"
> > Option "VertScrollDelta" "100"
> > Option "MinSpeed" "0.06"
> > Option "MaxSpeed" "0.12"
> > Option "AccelFactor" "0.0010"
> > Option "SHMConfig" "on"
> > Option "Repeater" "/dev/ps2mouse"
> > option "UpDownScrolling" "true"
> > EndSection
> > Section "Monitor"
> > Identifier "My Monitor"
> > HorizSync 31.5 - 57.0
> > VertRefresh 40-150
> > EndSection
> > Section "Device"
> > Identifier "Standard VGA"
> > VendorName "Unknown"
> > BoardName "Unknown"
> > Driver "vga"
> > EndSection
> > Section "Device"
> > Identifier "** S3 Savage (generic) [savage]"
> > Driver "savage"
> > EndSection
> > Section "Screen"
> > Identifier "Screen 1"
> > Device "** S3 Savage (generic) [savage]"
> > Monitor "My Monitor"
> > DefaultDepth 24
> > Subsection "Display"
> > Depth 8
> > Modes "1024x768" "800x600"
> > ViewPort 0 0
> > EndSubsection
> > Subsection "Display"
> > Depth 16
> > Modes "1024x768" "800x600"
> > ViewPort 0 0
> > EndSubsection
> > Subsection "Display"
> > Depth 24
> > Modes "1024x768" "800x600"
> > ViewPort 0 0
> > EndSubsection
> > EndSection
> > Section "ServerLayout"
> > Identifier "Simple Layout"
> > Screen "Screen 1"
> > InputDevice "Synaptic Mouse" "CorePointer"
> > InputDevice "Keyboard1" "CoreKeyboard"
> > EndSection
> > Section "DRI"
> > Mode 0666
> > EndSection
| Felix Kühling <fxkuehl@xxxxxx> http://fxk.de.vu |
| PGP Fingerprint: 6A3C 9566 5B30 DDED 73C3 B152 151C 5CC1 D888 E595 |
-------------------------------------------------------
This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
Project Admins to receive an Apple iPod Mini FREE for your judgement on
who ports your project to Linux PPC the best. Sponsored by IBM.
Deadline: Sept. 24. Go here: http://sf.net/ppc_contest.php
--
Previous Message by Thread:
Performance problem with radeon 9200SE PCI
I have a two-user PC (two screens, keyboard & mice). The idea
is to have two users being able to play 3D-games simultaneously.
(Kids are like that - if one play tuxracer then the other want the
same thing simultaneously . . .)
My setup works, but is a bit slow.
Hardware:
One matrox G550 AGP 1x, doing 1280x1024 16-bit. Performance is fine,
tuxracer is smooth at 1280x1024, I get 50fps in pinball,
glxgears gets 550fps at default size, or 80fps when maximized.
It is slower bit still ok with 24-bit color, but unstable when changing
resolution.
So I use it in 16-bit mode.
One radeon 9200SE PCI, doing 1280x1024 24-bit.
Tuxracer works, but is too choppy at 1280x1024. Of course I
can lower resolution, but I hope for more.
pinball gets 24fps, other games are also noticeably slower
glxgears gets 394fps, or 59fps maximized.
Cpu: AMD opteron 1.8GHz
Software:
I use linux, kernel 2.6.8.1 with the built-in DRM drivers for
radeon and matrox. The kernel is patched with the ruby patch
in order to support the dual keyboard setup. There are no
other kernel patches. I use debian testing with
some packages from unstable, my X is version 4.3.0.1
From glxinfo:
server glx vendor string: SGI
server gclient glx vendor string: SGI
client glx vendor string: SGI
client glx version string: 1.2
OpenGL vendor string: Tungsten Graphics, Inc.
OpenGL renderer string: Mesa DRI R200 20030328 x86/MMX+/3DNow!+/SSE2 TCL
OpenGL version string: 1.3 Mesa 6.1
I am hoping there is a way to get more performance from
the radeon card - the matrox is so old and still better? The reason
for choosing a radeon was a report that it actually works with
hw 3D as a _secondary_ card.
I don't know much about configuring for 3D performance.
I have seen some tips about setting the chipID. X identifies
it as chipID 5964. There are recommendations for
ovverriding with 0x514D to get better 3D performance, it makes
no difference at all for me.
Helge Hafting
-------------------------------------------------------
This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
Project Admins to receive an Apple iPod Mini FREE for your judgement on
who ports your project to Linux PPC the best. Sponsored by IBM.
Deadline: Sept. 24. Go here: http://sf.net/ppc_contest.php
--
Next Message by Thread:
Re: Performance problem with radeon 9200SE PCI
Roland Scheidegger wrote:
Helge Hafting wrote:
One radeon 9200SE PCI, doing 1280x1024 24-bit. Tuxracer works, but is
too choppy at 1280x1024. Of course I can lower resolution, but I
hope for more. pinball gets 24fps, other games are also noticeably
slower glxgears gets 394fps, or 59fps maximized.
Cpu: AMD opteron 1.8GHz
Well, the 9200SE is a pretty slow card to begin with (should still not
be slower than a G550 though).
I didn't expect a cheap card to be a speed daemon. :-) Better than G550
is good enough though. I bought a radeon card because radeon PCI is
reported
to actually work in a dual-seat DRI setup.
However, there are known performance problems with the pci based
R200-RV280 cards with the dri drivers. Unfortunately, no one knows why
:-(.
Ouch. Seems I'll have to follow the devel-list too then. I was hoping I
merely
had a bad default config.
I don't know much about configuring for 3D
performance. I have seen some tips about setting the chipID. X
identifies it as chipID 5964. There are recommendations for
ovverriding with 0x514D to get better 3D performance, it makes no
difference at all for me.
I don't think there's anything you could do, you might try newest
drivers from Mesa cvs. Not sure what version you're really using,
XFree86 4.3 and Mesa 6.1 typically aren't packaged together...
I use debian testing, with packages from unstable or experimental as
needed. I also
compile my own kernels.
Changing chipids will do nothing and should be avoided, it will only
help if the driver doesn't detect the card correctly if it's too new.
Thanks, I'll remove the chipid override then.
Helge Hafting
-------------------------------------------------------
This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
Project Admins to receive an Apple iPod Mini FREE for your judgement on
who ports your project to Linux PPC the best. Sponsored by IBM.
Deadline: Sept. 24. Go here: http://sf.net/ppc_contest.php
--
|
|