Go Back   nV News Forums > Linux Support Forums > NVIDIA Linux

Newegg Daily Deals

Reply
 
Thread Tools
Old 11-15-03, 01:32 AM   #1
squarooticus
 
Join Date: Nov 2003
Posts: 36
Default Slow PCI performance

In my previous system (dual Athlon, Tyan Tiger MP mobo), a GeForce2 MX/MX 400 PCI card was quick enough as my second display to easily play DVD's full screen at full framerate.

However, with my current system (dual Opteron, Tyan Thunder K8W mobo), the same card gives me absolutely abyssmal performance: it looks like it's in PIO mode, that's how slow it is. I can see a browser window repaint, for instance.

I'm still not sure if the problem originates with the kernel, the motherboard, or the nVidia drivers: I notice a kernel log message on startup indicating that it cannot assign resource regions 0 & 1 (memory-mapped I/O, apparently) of this card, but after some digging and adding printk's to various places, I determined that it is reassigning the card's memory-mapped I/O to a different region. (Unless this code is really broken, it verifies that the change actually "stuck" on the card by reading what it just wrote and verifying it; see drivers/pci/setup-res.c.)

At this point, I'm basically at a loss. There's no real way for me to tell whether this is a general problem with the PCI bus on this motherboard, or if the kernel doesn't have proper support for this PCI bus, or if the nVidia driver isn't actually using the memory-mapped regions. Is there a way for me to get this information so I can (perhaps) rule out the nVidia driver as the source of the problem?

Thanks,
Kyle

Last edited by squarooticus; 11-15-03 at 04:27 AM.
squarooticus is offline   Reply With Quote
Old 11-15-03, 12:50 PM   #2
squarooticus
 
Join Date: Nov 2003
Posts: 36
Default Did a lot more investigation

(EDIT: I don't know what happened to the log, so I'm attaching it in a later message. -krose)

I came up with two configurations that point out a problem either in the kernel or in the nVidia driver.

This message includes the first configuration.

The open source nv driver gives perfect 2D performance (i.e., full frame rate even in full-screen mode) on the PCI card with respect to playing DVD's with the same physical configuration, but with the AGP card unused by the software: essentially, I set up the second display with the nv driver and disabled the first display, and notwithstanding some other performance issues, the thing I care about (DVD playback) worked fine.

However, the nv driver doesn't seem to work with a dual head configuration, or at least won't drive the AGP card; besides, I want decent 3D acceleration on one of the displays so I can play games.

So, I naturally then tried to set up the AGP display with the nvidia driver and the PCI display with the nv driver. This didn't work at first: the system would freeze every time I tried to load X. I then changed the BIOS to start with the PCI VGA card instead of the AGP card, and got the log I am attaching to this message. The interesting part is this:

(WW) ****INVALID MEM ALLOCATION**** b: 0x40000000 e: 0x40ffffff correcting
(WW) ****INVALID MEM ALLOCATION**** b: 0x48000000 e: 0x4fffffff correcting

This error is associated with the *AGP card* when the PCI VGA display is the startup device: dmesg indicates that the AGP card's IOMEM is being remapped in this case, *not* the PCI card's! Note that I don't get this message about the PCI card when the AGP card is the master; I just get lousy performance.

Here's my XF86Config-4 for this attempted configuration. The log is attached. The second configuration is similar, and follows in the next message.

Section "Files"
#FontPath "unix/:7100" # local font server
# if the local font server has problems, we can fall back on these
FontPath "/usr/local/share/fonts/truetype"
#FontPath "/usr/local/share/fonts/Type1"
FontPath "/usr/local/share/fonts/misc"
FontPath "/usr/lib/X11/fonts/misc"
FontPath "/usr/lib/X11/fonts/cyrillic"
FontPath "/usr/lib/X11/fonts/75dpi/:unscaled"
FontPath "/usr/lib/X11/fonts/100dpi/:unscaled"
FontPath "/usr/lib/X11/fonts/Type1"
FontPath "/usr/lib/X11/fonts/Speedo"
FontPath "/usr/lib/X11/fonts/75dpi"
FontPath "/usr/lib/X11/fonts/100dpi"
EndSection

Section "Module"
#Load "GLcore"#
Load "bitmap"
Load "dbe"
Load "ddc"
#Load "dri"#
Load "extmod"
Load "freetype"
Load "glx"
Load "int10"
Load "pex5"
Load "record"
Load "speedo"
Load "type1"
Load "vbe"
Load "xie"
EndSection

Section "InputDevice"
Identifier "Generic Keyboard"
Driver "keyboard"
Option "CoreKeyboard"
Option "XkbRules" "xfree86"
Option "XkbModel" "pc104"
Option "XkbLayout" "us"
EndSection

Section "InputDevice"
Identifier "Configured Mouse"
Driver "mouse"
Option "CorePointer"
Option "Device" "/dev/input/mice"
Option "Protocol" "ExplorerPS/2"
Option "Emulate3Buttons" "no"
Option "ZAxisMapping" "5 6"
Option "Buttons" "6"
EndSection

Section "Device"
Identifier "MSI FX5600"
Driver "nvidia"
Option "NoLogo" "true"
Option "NvAGP" "3"
Option "ConnectedMonitor" "DFP"
Option "UseEdidFreqs"
Option "FlatPanelProperties" "Scaling = scaled, Dithering = default"
Option "DigitalVibrance" "100"
BusID "PCI:5:0:0"
EndSection

Section "Device"
Identifier "XFX GeForce FX 5200"
Driver "nv"
Option "HWCursor" "false"
BusID "PCI:1:10:0"
EndSection

Section "Monitor"
Identifier "NEC LCD1860NX"
HorizSync 31-81
VertRefresh 56-60
Option "DPMS"
EndSection

Section "Monitor"
Identifier "ViewSonic VP181b"
HorizSync 30-94
VertRefresh 50-75
Option "DPMS"
EndSection

Section "Screen"
Identifier "Screen 1"
Device "MSI FX5600"
Monitor "ViewSonic VP181b"
DefaultDepth 24
SubSection "Display"
Depth 8
Modes "1280x1024" "1024x768"
EndSubSection
SubSection "Display"
Depth 16
Modes "1280x1024" "1024x768"
EndSubSection
SubSection "Display"
Depth 24
Modes "1280x1024" "1024x768"
EndSubSection
EndSection

Section "Screen"
Identifier "Screen 2"
Device "XFX GeForce FX 5200"
Monitor "NEC LCD1860NX"
DefaultDepth 24
SubSection "Display"
Depth 8
Modes "1280x1024" "1024x768"
EndSubSection
SubSection "Display"
Depth 16
Modes "1280x1024" "1024x768"
EndSubSection
SubSection "Display"
Depth 24
Modes "1280x1024" "1024x768"
EndSubSection
EndSection

Section "ServerLayout"
Identifier "Default Layout"
Screen "Screen 1"
Screen "Screen 2" RightOf "Screen 1"
InputDevice "Generic Keyboard"
InputDevice "Configured Mouse" "CorePointer"
EndSection

Section "ServerFlags"
Option "Xinerama" "false"
Option "DontZap" "true"
EndSection

Section "DRI"
Mode 0666
EndSection

Last edited by squarooticus; 11-15-03 at 01:18 PM.
squarooticus is offline   Reply With Quote
Old 11-15-03, 12:51 PM   #3
squarooticus
 
Join Date: Nov 2003
Posts: 36
Default 2nd configuration

In this configuration, I decided to try the nvidia driver on both displays, but keeping the PCI VGA card as the master. I get the same error in the attached log. Here's the configuration.

Section "Files"
#FontPath "unix/:7100" # local font server
# if the local font server has problems, we can fall back on these
FontPath "/usr/local/share/fonts/truetype"
#FontPath "/usr/local/share/fonts/Type1"
FontPath "/usr/local/share/fonts/misc"
FontPath "/usr/lib/X11/fonts/misc"
FontPath "/usr/lib/X11/fonts/cyrillic"
FontPath "/usr/lib/X11/fonts/75dpi/:unscaled"
FontPath "/usr/lib/X11/fonts/100dpi/:unscaled"
FontPath "/usr/lib/X11/fonts/Type1"
FontPath "/usr/lib/X11/fonts/Speedo"
FontPath "/usr/lib/X11/fonts/75dpi"
FontPath "/usr/lib/X11/fonts/100dpi"
EndSection

Section "Module"
#Load "GLcore"#
Load "bitmap"
Load "dbe"
Load "ddc"
#Load "dri"#
Load "extmod"
Load "freetype"
Load "glx"
Load "int10"
Load "pex5"
Load "record"
Load "speedo"
Load "type1"
Load "vbe"
Load "xie"
EndSection

Section "InputDevice"
Identifier "Generic Keyboard"
Driver "keyboard"
Option "CoreKeyboard"
Option "XkbRules" "xfree86"
Option "XkbModel" "pc104"
Option "XkbLayout" "us"
EndSection

Section "InputDevice"
Identifier "Configured Mouse"
Driver "mouse"
Option "CorePointer"
Option "Device" "/dev/input/mice"
Option "Protocol" "ExplorerPS/2"
Option "Emulate3Buttons" "no"
Option "ZAxisMapping" "5 6"
Option "Buttons" "6"
EndSection

Section "Device"
Identifier "MSI FX5600"
Driver "nvidia"
Option "NoLogo" "true"
Option "NvAGP" "3"
Option "ConnectedMonitor" "DFP"
Option "UseEdidFreqs"
Option "FlatPanelProperties" "Scaling = scaled, Dithering = default"
Option "DigitalVibrance" "100"
BusID "PCI:5:0:0"
EndSection

Section "Device"
Identifier "XFX GeForce FX 5200"
Driver "nvidia"
Option "NoLogo" "true"
Option "NvAGP" "0"
#Option "UseInt10Module" "true"
BusID "PCI:1:10:0"
EndSection

Section "Monitor"
Identifier "NEC LCD1860NX"
HorizSync 31-81
VertRefresh 56-60
Option "DPMS"
EndSection

Section "Monitor"
Identifier "ViewSonic VP181b"
HorizSync 30-94
VertRefresh 50-75
Option "DPMS"
EndSection

Section "Screen"
Identifier "Screen 1"
Device "MSI FX5600"
Monitor "ViewSonic VP181b"
DefaultDepth 24
SubSection "Display"
Depth 8
Modes "1280x1024" "1024x768"
EndSubSection
SubSection "Display"
Depth 16
Modes "1280x1024" "1024x768"
EndSubSection
SubSection "Display"
Depth 24
Modes "1280x1024" "1024x768"
EndSubSection
EndSection

Section "Screen"
Identifier "Screen 2"
Device "XFX GeForce FX 5200"
Monitor "NEC LCD1860NX"
DefaultDepth 24
SubSection "Display"
Depth 8
Modes "1280x1024" "1024x768"
EndSubSection
SubSection "Display"
Depth 16
Modes "1280x1024" "1024x768"
EndSubSection
SubSection "Display"
Depth 24
Modes "1280x1024" "1024x768"
EndSubSection
EndSection

Section "ServerLayout"
Identifier "Default Layout"
Screen "Screen 1"
Screen "Screen 2" RightOf "Screen 1"
InputDevice "Generic Keyboard"
InputDevice "Configured Mouse" "CorePointer"
EndSection

Section "ServerFlags"
Option "Xinerama" "false"
Option "DontZap" "true"
EndSection

Section "DRI"
Mode 0666
EndSection
Attached Files
File Type: log x.3.log (23.4 KB, 98 views)
squarooticus is offline   Reply With Quote
Old 11-15-03, 01:17 PM   #4
squarooticus
 
Join Date: Nov 2003
Posts: 36
Default Educated guess about what's happening

Given what I've witnessed (good 2D performance with nv, and nvidia driver's inability to cope with IOMEM remaps to 0x40000000), my educated guess is that either the nVidia driver or the nVidia kernel module is at fault here. If there were a way to get more interesting debugging information out of the driver or the kernel module, it might be possible for me to further debug this; but as it is, I am very frustrated with the lack of transparency into the driver.

My configuration (more specifically) is:

Tyan Thunder K8W 2885 with AMD-8151 AGP bridge, AMD-8111 PCI bridge, and AMD-8131 PCI-X bridge
Dual Opteron 244
nVidia NV31 GeForce FX 5600 AGP
nVidia NV34 GeForce FX 5200 PCI
Linux 2.6.0-test9-mm3

Here are the relevant dmesg lines from my normal config (AGP master, nvidia driver on both screens):

PCI: Cannot allocate resource region 0 of device 0000:02:08.0
PCI: Cannot allocate resource region 1 of device 0000:02:08.0
Linux agpgart interface v0.100 (c) Dave Jones
agpgart: Detected AMD 8151 AGP Bridge rev B2
agpgart: Maximum main memory to use for agp memory: 941M
agpgart: AGP aperture is 256M @ 0xe0000000
nvidia: no version for "struct_module" found: kernel tainted.
nvidia: no version magic, tainting kernel.
nvidia: module license 'NVIDIA' taints kernel.
1: nvidia: loading NVIDIA Linux x86 nvidia.o Kernel Module 1.0-4496 Wed Jul 16 19:03:09 PDT 2003
agpgart: Found an AGP 3.0 compliant device at 0000:05:00.0.
agpgart: Putting AGP V3 device at 0000:05:00.0 into 8x mode
agpgart: Putting AGP V3 device at 0000:06:00.0 into 8x mode

And here's my lspci output, with -v for the nVidia stuff:

00:06.0 PCI bridge: Advanced Micro Devices [AMD] AMD-8111 PCI (rev 07)
00:07.0 ISA bridge: Advanced Micro Devices [AMD] AMD-8111 LPC (rev 05)
00:07.1 IDE interface: Advanced Micro Devices [AMD] AMD-8111 IDE (rev 03)
00:07.2 SMBus: Advanced Micro Devices [AMD] AMD-8111 SMBus 2.0 (rev 02)
00:07.3 Bridge: Advanced Micro Devices [AMD] AMD-8111 ACPI (rev 05)
00:0a.0 PCI bridge: Advanced Micro Devices [AMD] AMD-8131 PCI-X Bridge (rev 12)
00:0a.1 PIC: Advanced Micro Devices [AMD] AMD-8131 PCI-X APIC (rev 01)
00:0b.0 PCI bridge: Advanced Micro Devices [AMD] AMD-8131 PCI-X Bridge (rev 12)
00:0b.1 PIC: Advanced Micro Devices [AMD] AMD-8131 PCI-X APIC (rev 01)
00:18.0 Host bridge: Advanced Micro Devices [AMD] K8 NorthBridge
00:18.1 Host bridge: Advanced Micro Devices [AMD] K8 NorthBridge
00:18.2 Host bridge: Advanced Micro Devices [AMD] K8 NorthBridge
00:18.3 Host bridge: Advanced Micro Devices [AMD] K8 NorthBridge
00:19.0 Host bridge: Advanced Micro Devices [AMD] K8 NorthBridge
00:19.1 Host bridge: Advanced Micro Devices [AMD] K8 NorthBridge
00:19.2 Host bridge: Advanced Micro Devices [AMD] K8 NorthBridge
00:19.3 Host bridge: Advanced Micro Devices [AMD] K8 NorthBridge
02:07.0 Multimedia audio controller: Creative Labs SB Live! EMU10k1 (rev 07)
02:07.1 Input device controller: Creative Labs SB Live! MIDI/Game Port (rev 07)

02:08.0 VGA compatible controller: nVidia Corporation NV34 [GeForce FX 5200] (rev a1)
Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B-
Status: Cap+ 66Mhz+ UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR-
Latency: 248 (1250ns min, 250ns max)
Interrupt: pin A routed to IRQ 27
Region 0: Memory at fb000000 (32-bit, non-prefetchable) [size=16M]
Region 1: Memory at b0000000 (32-bit, prefetchable) [size=128M]
Expansion ROM at fffe0000 [disabled] [size=128K]
Capabilities: <available only to root>

02:09.0 Ethernet controller: Broadcom Corporation NetXtreme BCM5703 Gigabit Ethernet (rev 02)
03:00.0 USB Controller: Advanced Micro Devices [AMD] AMD-8111 USB (rev 0b)
03:00.1 USB Controller: Advanced Micro Devices [AMD] AMD-8111 USB (rev 0b)
03:0a.0 PCI bridge: Hint Corp HB6 Universal PCI-PCI bridge (non-transparent mode) (rev 15)
03:0b.0 Unknown mass storage controller: CMD Technology Inc: Unknown device 3114 (rev 02)
03:0c.0 FireWire (IEEE 1394): Texas Instruments TSB43AB22/A IEEE-1394a-2000 Controller (PHY/Link)
04:08.0 USB Controller: NEC Corporation USB (rev 41)
04:08.1 USB Controller: NEC Corporation USB (rev 41)
04:08.2 USB Controller: NEC Corporation USB 2.0 (rev 02)
04:09.0 FireWire (IEEE 1394): VIA Technologies, Inc. IEEE 1394 Host Controller (rev 46)
05:00.0 Host bridge: Advanced Micro Devices [AMD] AMD-8151 System Controller (rev 13)
05:01.0 PCI bridge: Advanced Micro Devices [AMD] AMD-8151 AGP Bridge (rev 13)

06:00.0 VGA compatible controller: nVidia Corporation NV31 [GeForce FX 5600] (rev a1)
Subsystem: Micro-Star International Co., Ltd.: Unknown device 9123
Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B-
Status: Cap+ 66Mhz+ UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR-
Latency: 248 (1250ns min, 250ns max)
Interrupt: pin A routed to IRQ 16
Region 0: Memory at fd000000 (32-bit, non-prefetchable) [size=16M]
Region 1: Memory at c0000000 (32-bit, prefetchable) [size=256M]
Expansion ROM at feae0000 [disabled] [size=128K]
Capabilities: <available only to root>

AND, here's the output of /proc/driver/nvidia/card/*:

Model: GeForce FX 5200
IRQ: 27
Video BIOS: 04.34.20.27.00
Card Type: PCI

Model: GeForce FX 5600
IRQ: 16
Video BIOS: 04.31.20.26.02
Card Type: AGP

If you're going to claim Linux compatibility but not release the driver source to the community, you need to provide better support: I've seen far too many eerily-similar complaints about the 2D performance to justify going this long without an investigation and fix. I am going to put a hold on my own purchases until I see both a fix for this particular problem and a general improvement in nVidia's attitude toward Linux users.

Cheers,
Kyle

Last edited by squarooticus; 11-15-03 at 01:30 PM.
squarooticus is offline   Reply With Quote
Old 11-15-03, 01:20 PM   #5
squarooticus
 
Join Date: Nov 2003
Posts: 36
Default X.1.log

Here's the disappeared log from the second message.
Attached Files
File Type: log x.1.log (27.1 KB, 101 views)
squarooticus is offline   Reply With Quote
Old 11-17-03, 12:19 AM   #6
squarooticus
 
Join Date: Nov 2003
Posts: 36
Default bump

I hope someone from nVidia has a comment or a potential solution.
squarooticus is offline   Reply With Quote
Old 11-17-03, 10:54 PM   #7
squarooticus
 
Join Date: Nov 2003
Posts: 36
Default Aha! More information, and some success

Unfortunately, it's with the open source driver.

I managed to get DVD playback working on both screens in dual head mode with the open source nv driver from XFree86-4.3. It takes a bit of massaging: whenever I first boot, I need to initialize both cards by bringing up X with the nvidia proprietary driver, but then I just quit X and reload with the open source driver, and all is well: fast 2D performance on both screens.

I think this pretty much settles it: there's a problem with the nVidia driver, not with the hardware or with the kernel.

Cheers,
Kyle
squarooticus is offline   Reply With Quote
Old 08-12-04, 11:22 PM   #8
squarooticus
 
Join Date: Nov 2003
Posts: 36
Default Re: Slow PCI performance

This post deleted because I am an idiot.
__________________
Tyan Thunder K8W 2885
Dual Opteron 244
XFX 6600GT AGP
4 GB PC 2700 Reg ECC DDR SDRAM

Last edited by squarooticus; 08-12-04 at 11:56 PM.
squarooticus is offline   Reply With Quote

Old 12-05-04, 03:59 PM   #9
ifaxsoln
Registered User
 
Join Date: Nov 2004
Posts: 21
Default Re: Slow PCI performance

Quote:
Originally Posted by squarooticus
This post deleted because I am an idiot.
Care to elaborate? I'm seeing the same problem with the nvidia driver (abysmal X performance) with the same motherboard ...

I guess I'm an idiot too, but I'd like to know why!

-d
ifaxsoln is offline   Reply With Quote
Old 12-05-04, 05:17 PM   #10
zander
NVIDIA Corporation
 
zander's Avatar
 
Join Date: Aug 2002
Posts: 3,740
Default Re: Slow PCI performance

@ifaxsoln: have you tried: http://www.nvnews.net/vbulletin/show...15&postcount=2?
zander is offline   Reply With Quote
Old 12-05-04, 05:40 PM   #11
ifaxsoln
Registered User
 
Join Date: Nov 2004
Posts: 21
Default Re: Slow PCI performance

I didn't find that in my searching, no. I'm using the x86_64 drivers, but I applied this fix and IT WORKED. Performance is _much_ improved.

My only remaining problem is the inability to recognize both GPUs on a Quadro NVS 400, ... I'll start a separate thread for that though since I've hijacked this one badly already.
ifaxsoln is offline   Reply With Quote
Reply


Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


Similar Threads
Thread Thread Starter Forum Replies Last Post
Boost Your Performance Goals 10x ' This Week on inside* Publications News Archived News Items 0 05-13-12 06:20 PM
NVIDIA Unleashes the GeForce GTX 670 Graphics Card ' Performance Perfected (WCCFTECH) News GeForce GTX 670 Reviews 0 05-10-12 09:40 AM
My UT2003 Tweak Guide DXnfiniteFX Gaming Central 48 10-31-02 12:59 AM
X Failed to load NVdriver c00lr4c3r NVIDIA Linux 13 10-22-02 02:44 PM
Need help to get the X to work on my Acer TravelMate 630 knchee NVIDIA Linux 16 09-19-02 11:16 PM

All times are GMT -5. The time now is 11:11 AM.


Powered by vBulletin® Version 3.7.1
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Copyright 1998 - 2014, nV News.