|
Replies:
18
-
Last Post:
Sep 21, 2008 10:00 PM
by: Ric Moore
|
|
|
|
|
|
|
Performance issue on quad-core machine
Posted:
Aug 24, 2008 1:51 AM
|
|
|
Hi there,
I'm trying to track down a performance problem with my Wonderland system. The symptom is that if I stand in place and look around, the system frequently freezes for long periods--ten seconds or more is not uncommon--before it wakes up and renders. The delays seem to get longer as I move around in the system and look at things. After a while I'm seeing delays of 30 seconds. This is after setting the max frame rate to 15.
I've been running the pre-built Wonderland 0.3.0 from Linux (lg3d-wonderland+voice-0-3-0-linux-i686-20080110.zip). I just tried upgrading to the latest release (wonderland-0-4-0-linux-i686.zip) and that didn't seem to help. (It did have the effect that my shared applications stopped working, but let me set that aside for now - I'd rather fix the performance problem.)
I have everything (bridge, server, SMC, and client) running on the same box. System stats:
CPU: AMD Phenom(tm) 9850 Quad-Core Processor, 2.5 GHz
Memory: 1945568 kB
Graphics (on-board): NVIDIA GPU GeForce 6150SE nForce 430 (C61)
Graphics driver: Kernel module and xorg driver downloaded from nvidia.com, version 173.14.09
Linux Kernel: 2.6.26 with SMP support, booting with "nolapic" option
OS: Debian "etch" (stable) for the voice bridge and the server, Debian "sid" (unstable) in a chroot system for the application server and the client (since they apparently require GLIBC 2.4 to work correctly)
JVM: Sun Java SE Runtime Environment (build 1.6.0_07-b06), HotSpot Server VM (build 10.0-b23, mixed mode)
Running "glxinfo | grep rendering" yields "direct rendering: Yes". Running "glxgears -printfps" tells me I'm getting over 295 FPS at full-screen. So I don't think I have a problem with the 3-D acceleration under X. However, if I run "top" and induce the freeze-up, I see that the Xorg process is chewing up the CPU. This makes me wonder whether Java3D is failing to take advantage of the direct rendering available. I've checked through the console output from the client, but I'm not finding any complaints that seem relevant to the issue.
I'm running the Wonderland client inside a chroot system which has the newer GLIBC, because that's apparently necessary to make the sound work. But I've run it on the stable system, and apart from the lack of sound, I get similar results. So I don't think that's to blame for the poor performance.
There is nothing else of note running on this box. It's a dedicated Wonderland system.
my.run.properties (comments deleted, hostname concealed):
sgs.server=127.0.0.1
sgs.port=1139
wonderland.local.hostAddress=10.1.10.4
voicebridge.enabled=true
voicebridge.server.public.address=[[ my system's hostname ]]
voicebridge.server.public.sip.port=5060
voicebridge.first.rtp.port=5061
voicebridge.last.rtp.port=5262
wonderland.appshare.minPort=44400
wonderland.appshare.maxPort=44500
wonderland.appshare.hostName=[[ my system's hostname ]]
wonderland.useLocalArt=false
Any hints on how to make Wonderland's performance more like glxgears's would be most welcome.
Thanks,
Greg
|
|
|
|
|
|
|
Re: Performance issue on quad-core machine
Posted:
Aug 24, 2008 5:42 AM
in response to: lord_max
|
 |
Helpful |
|
|
I was just doing this on Ubuntu machines -- I installed the driver from nVidia and ran glxgears and got ~250 FPS. I then went to System -> Hardware Drivers and enabled the nVidia driver and re-ran glxgears and got ~8000 FPS (but not at full-screen). The Wonderland client didn't work (I actually got a gray screen) until I enabled the hardware drivers. Don't know if you are running into the same problem.
Also, your logs (e.g. wonderland-client.log, wonderland-sgs.log) should give you information about any driver issues. You are welcome to post them here and we can have a look to see if anything is obvious.
|
|
|
|
|
|
|
|
|
Re: Performance issue on quad-core machine
Posted:
Aug 24, 2008 8:46 PM
in response to: jslott
|
|
|
Hi, thanks for responding.
Since I'm running Debian, I don't see the same menu options as you do. If I'm going to change my X configuration, I'm assuming I'll have to do it by hand. If you wouldn't mind, could you 1) disable the hardware nVidia driver, 2) make a copy of /etc/X11/xorg.conf, 3) re-enable the driver, and 4) post a diff of the old xorg.conf and the new one?
It would also be nice to know if your xorg.conf contains explicit DRI support. Mine does not. I think I read somewhere that nVidia's driver doesn't use it, but instead, talks directly to the kernel module. I could have that wrong though. I'll try playing with that, just in case.
I do have something called System Tools / NVIDIA X Server Settings. It tells me, for example, that I have driver version 173.14.09, X Server version 11.0 and vendor version 7.1.1, NV-Control version 1.16, and a GeForce 6150SE nForce 430.
In that control panel, I have a "Sync to VBlank" option under Video Texture Adaptor (enabled) and another "Sync to VBlank" under Video Blitter Adaptor Settings (disabled). Under OpenGL settings I again have "Sync to VBlank" (disabled) and "Allow Flipping" (enabled), and an image settings slider, currently set at "Quality" - I can move it down to "High Quality", or up to "Performance" or "High Performance". I will try tinkering with some of these, too, and see if they help.
Under OpenGL/GLX Information I have a read-only panel which, sadly, doesn't allow me to easily copy and paste everything. It says "Direct Rendering: Yes", and has a few long lists of stuff - GLX Extensions, Server GLX Extensions, Client GLX Extensions, OpenGL Extensions. Let me know if you'd find any of these interesting; they're all things like "GLX_SGI_swap_control" and "GLX_NV_float_buffer", presumably supported hardware acceleration features. Under OpenGL it says "Renderer: GeForce 6150SE nForce 430/PCI/SSE2/3DNOW!".
I don't have either of the log files you mentioned - I have sipMessages.log, bridge.log, and nistStack.log. Perhaps this is because I'm running everything in separate windows under "screen", and the output is simply going to the console.
Here's the complete console output from wonderland-client.sh, for a three-minute session in which I connected, ran around a bit, and saw it pause for several seconds:
Buildfile: /data/app/lg3d-wonderland/ant/run.xml
run-client:
[java] Aug 24, 2008 7:51:19 PM org.jdesktop.lg3d.wonderland.LocalControl getLocalControlFile
[java] WARNING: Couldn't get local host address
[java] java.net.UnknownHostException: coin: coin
[java] at java.net.InetAddress.getLocalHost(InetAddress.java:1353)
[java] at org.jdesktop.lg3d.wonderland.LocalControl.getLocalControlFile(LocalControl.java:302)
[java] at org.jdesktop.lg3d.wonderland.LocalControl.connectToLocalInstance(LocalControl.java:249)
[java] at org.jdesktop.lg3d.wonderland.LocalControl.isFirstInstance(LocalControl.java:114)
[java] at org.jdesktop.lg3d.wonderland.Main.setupLocalControl(Main.java:2016)
[java] at org.jdesktop.lg3d.wonderland.Main.main(Main.java:1949)
[java] Aug 24, 2008 7:51:21 PM org.jdesktop.lg3d.wonderland.config.common.WonderlandConfigUtil getInputStream
[java] INFO: Loading wonderland configuration file: /home/samvimes/.wonderland/config/WonderlandClientConfig.xml
[java] Aug 24, 2008 7:51:22 PM org.jdesktop.lg3d.wonderland.config.common.WonderlandConfigUtil getInputStream
[java] WARNING: Error loading config file /home/samvimes/.wonderland/config//apps-linux/apps.xml : /home/samvimes/.wonderland/config/apps-linux/apps.xml (No such file or directory)
[java] Aug 24, 2008 7:51:22 PM org.jdesktop.lg3d.wonderland.config.common.WonderlandConfigUtil getInputStream
[java] INFO: Loading wonderland configuration file: /data/app/lg3d-wonderland/config/apps-linux/apps.xml
[java] Aug 24, 2008 7:51:22 PM org.jdesktop.lg3d.wonderland.config.common.WonderlandConfigUtil getInputStream
[java] WARNING: Error loading config file /home/samvimes/.wonderland/config/WonderlandConfig.xml : /home/samvimes/.wonderland/config/WonderlandConfig.xml (No such file or directory)
[java] Aug 24, 2008 7:51:22 PM org.jdesktop.lg3d.wonderland.config.common.WonderlandConfigUtil getInputStream
[java] INFO: Loading wonderland configuration file: /data/app/lg3d-wonderland/config/WonderlandConfig.xml
[java] Aug 24, 2008 7:51:23 PM org.jdesktop.lg3d.wonderland.config.common.WonderlandConfigUtil getInputStream
[java] INFO: Loading wonderland configuration file: /home/samvimes/.wonderland/config/AvatarClientConfig.xml
[java] 3D [dev] 1.5.3-build2-experimental 13 Aug 2008 16:42:15 PDT
[java] Java 3D: Cache AutoCompute Bounds, accelerates getBounds() enabled
[java] CameraManager.changeCameraMode(), current = ATTACHED
[java] Aug 24, 2008 7:51:23 PM org.jdesktop.lg3d.wonderland.appshare.MasterLauncher launchExternal
[java] WARNING: launch external command: [/data/app/lg3d-wonderland/bin/ffsetup, ./data/Wonderland/test/appshare]
[java] Aug 24, 2008 7:51:23 PM org.jdesktop.lg3d.wonderland.Main$ffsetupReporter output
[java] WARNING: Output for ffsetup: /bin/rm: cannot remove directory `/home/samvimes/.mozilla/firefox/giuw7do4.WonderlandFirefoxTeamRoomCenter/chrome': No such file or directory
[java] Aug 24, 2008 7:51:24 PM org.jdesktop.lg3d.wonderland.config.common.WonderlandConfigUtil getInputStream
[java] WARNING: Error loading config file /home/samvimes/.wonderland/config/WonderlandViewConfig.xml : /home/samvimes/.wonderland/config/WonderlandViewConfig.xml (No such file or directory)
[java] Aug 24, 2008 7:51:24 PM org.jdesktop.lg3d.wonderland.config.common.WonderlandConfigUtil getInputStream
[java] WARNING: Error loading config file /data/app/lg3d-wonderland/config/WonderlandViewConfig.xml : /data/app/lg3d-wonderland/config/WonderlandViewConfig.xml (No such file or directory)
[java] (<unknown>:2214): Gtk-WARNING **: Attempting to add a widget with type GtkButton to a GtkComboBoxEntry (need an instance of GtkEntry or of a subclass)
[java] (<unknown>:2214): Gtk-CRITICAL **: gtk_widget_realize: assertion `GTK_WIDGET_ANCHORED (widget) || GTK_IS_INVISIBLE (widget)' failed
[java] (<unknown>:2214): Gtk-CRITICAL **: gtk_paint_box: assertion `style->depth == gdk_drawable_get_depth (window)' failed
[java] (<unknown>:2214): Gtk-CRITICAL **: gtk_paint_box: assertion `style->depth == gdk_drawable_get_depth (window)' failed
[java] (<unknown>:2214): Gtk-CRITICAL **: gtk_paint_box: assertion `style->depth == gdk_drawable_get_depth (window)' failed
[java] (<unknown>:2214): Gtk-CRITICAL **: gtk_paint_box: assertion `style->depth == gdk_drawable_get_depth (window)' failed
[java] Aug 24, 2008 7:52:14 PM org.jdesktop.lg3d.wonderland.scenemanager.AssetDB setDBSystemDir
[java] FINE: AssetDB SystemDir = /home/samvimes/.wonderland/
[java] Aug 24, 2008 7:52:15 PM org.jdesktop.lg3d.wonderland.darkstar.client.ChannelController initCommunications
[java] INFO: Phone number is
[java] Aug 24, 2008 7:52:15 PM org.jdesktop.lg3d.wonderland.darkstar.client.ChannelController$WonderlandClientListener receivedMessage
[java] CONFIG: Checking protocol versions
[java] Aug 24, 2008 7:52:15 PM org.jdesktop.lg3d.wonderland.darkstar.client.ChannelController$WonderlandClientListener receivedMessage
[java] CONFIG: Server version 1.7_0
[java] Aug 24, 2008 7:52:15 PM org.jdesktop.lg3d.wonderland.darkstar.client.ChannelController$WonderlandClientListener receivedMessage
[java] CONFIG: Local version 1.7_0
[java] Aug 24, 2008 7:52:15 PM org.jdesktop.lg3d.wonderland.Main$71 run
[java] WARNING: Server shared apps are supported, number of apps = 7
[java] Aug 24, 2008 7:52:15 PM org.jdesktop.lg3d.wonderland.config.common.WonderlandConfigUtil getInputStream
[java] WARNING: Error loading config file /home/samvimes/.wonderland/config/InputConfig.xml : /home/samvimes/.wonderland/config/InputConfig.xml (No such file or directory)
[java] Aug 24, 2008 7:52:15 PM org.jdesktop.lg3d.wonderland.config.common.WonderlandConfigUtil getInputStream
[java] WARNING: Error loading config file /data/app/lg3d-wonderland/config/InputConfig.xml : /data/app/lg3d-wonderland/config/InputConfig.xml (No such file or directory)
[java] Aug 24, 2008 7:52:15 PM org.jdesktop.lg3d.wonderland.config.common.WonderlandConfigUtil getInputStream
[java] INFO: Loading wonderland configuration file: /home/samvimes/.wonderland/config/AvatarModelConfig.xml
[java] Aug 24, 2008 7:52:15 PM org.jdesktop.lg3d.wonderland.scenemanager.avatar.AvatarConfigurator getDefaultAvatarInfo
[java] INFO: Loading default avatar : [type=FEMALE],[bodyParts=[Eye_B, Eye_Brown_R, Eye_B, Eye_Brown_L, Glasses, Glasses2, Hair_, Hair_2, Hand_L, Hand_L2, Hand_R, Hand_R2, Head_, Head_1, UpperBody_, UpperBody_1]],[sharedMaterials=[SKin_BlackShader, EyesBrownShader]],[materialColors=[ShirtShader, #fd9911, PantsShader, #e867ae, ShoesShader, #fd62e2, GlassesShader, #045dd9, HairShader, #f9fb30]] [java] Aug 24, 2008 7:52:15 PM org.jdesktop.lg3d.wonderland.darkstar.client.ChannelController connectSoftphone
[java] INFO: Sending message to server asking for a voice bridge
[java] Aug 24, 2008 7:52:15 PM org.jdesktop.lg3d.wonderland.config.common.WonderlandConfigUtil writeUserConfig
[java] INFO: Writing wonderland USER configuration file: /home/samvimes/.wonderland/config/AvatarClientConfig.xml
[java] Aug 24, 2008 7:52:15 PM org.jdesktop.lg3d.wonderland.config.common.WonderlandConfigUtil getInputStream
[java] INFO: Loading wonderland configuration file: /data/app/lg3d-wonderland/config/PlacemarkConfig.xml
[java] Aug 24, 2008 7:52:15 PM org.jdesktop.lg3d.wonderland.config.common.WonderlandConfigUtil getInputStream
[java] WARNING: Error loading config file /home/samvimes/.wonderland/config/PlacemarkConfig-127.0.0.1-1139.xml : /home/samvimes/.wonderland/config/PlacemarkConfig-127.0.0.1-1139.xml (No such file or directory)
[java] Aug 24, 2008 7:52:16 PM com.sun.labs.mpk20.phone.client.cell.PhoneCell setup
[java] WARNING: Phone 1 locked is false
[java] Aug 24, 2008 7:52:23 PM com.sun.labs.mpk20.phone.client.cell.PhoneCell setup
[java] WARNING: Phone 2 locked is false
[java] TODO implement AppWindowGraphics2DApp.ensureIsAboveFloorAug 24, 2008 7:52:24 PM com.sun.labs.mpk20.phone.client.cell.PhoneCell setup
[java] WARNING: Phone 3 locked is false
[java] [SIP] Aug 24 19:52:25.824 Arguments: -mc -u samvimes -r 10.1.10.4;sip-stun:5060 -stun 10.1.10.4:5060 -t 10 -l 10.1.10.4 -sampleRate 16000 -channels 2 -transmitSampleRate 16000 -transmitChannels 1
[java] [SIP] Aug 24 19:52:26.57 Setting local IP Address to 10.1.10.4
[java] [SIP] Aug 24 19:52:26.108 Sip Communicator built on Thu Aug 7 2008 12:35 PM
[java] [SIP] Aug 24 19:52:26.108 Running java version 1.6.0_07
[java] [SIP] Aug 24 19:52:26.108 OS Name = Linux, OS Arch = i386, OS Version = 2.6.26coin1
[java] [SIP err] Aug 24, 2008 7:52:26 PM com.sun.stun.NetworkAddressManager getLocalHostFromPreferredAddress
[java] [SIP err] INFO: Using specified local address /10.1.10.4
[java] [SIP err] Aug 24, 2008 7:52:26 PM com.sun.stun.NetworkAddressManager getLocalHost
[java] [SIP err] INFO: Using preferred address 10.1.10.4
[java] [SIP err] Aug 24, 2008 7:52:26 PM com.sun.stun.NetworkAddressManager showDefaultAddress
[java] [SIP err] INFO: If localHost had not been specified, 10.1.10.4 would have been chosen by using STUN.
[java] [SIP] Aug 24 19:52:26.851 Log file is /tmp/sipCommunicator.samvimes_46903.log
[java] [SIP] Aug 24 19:52:26.852 SIP Server log file is /tmp/sipServer.samvimes_46902.log
[java] [SIP] Aug 24 19:52:27.6 log level is 3
[java] [SIP] Aug 24 19:52:27.80 Starting thread to watch for MC heartbeat.
[java] [SIP err] Aug 24, 2008 7:52:27 PM com.sun.stun.StunClient run
[java] [SIP err] INFO: using STUN server /10.1.10.4:5060
[java] [SIP] Aug 24 19:52:27.281 private address is 10.1.10.4:5070 publicIsa is /10.1.10.4:5070
[java] [SIP] SipCommunicator Public Address is 'sip:samvimes@10.1.10.4:5070'
[java] Ask server to call softphone: sip:samvimes@10.1.10.4:5070 position (50.0, 1.1400006, 50.0) direction (0.0, 0.0, -1.0)
[java] [SIP] Aug 24 19:52:27.292 Registering with registrar 10.1.10.4:5060 expiration 10
[java] [SIP err] Aug 24, 2008 7:52:27 PM com.sun.stun.StunClient run
[java] [SIP err] INFO: using STUN server /10.1.10.4:5072
[java] [SIP] Aug 24 19:52:27.653 generateSdp: Our isa /10.1.10.4:43586
[java] [SIP err] Aug 24, 2008 7:52:27 PM com.sun.stun.StunClient run
[java] [SIP err] INFO: using STUN server /10.1.10.4:5073
[java] [SIP] Aug 24 19:52:27.656 rtcp address is /10.1.10.4:43587
[java] [SIP err] native name is native/libMediaFramework.so
[java] [SIP] Aug 24 19:52:27.761 Initializing audio driver to 16000/2 bufferSize 10240
[java] [SIP] Aug 24 19:52:27.764 Using speaker preference plughw:0,0 NVidia [HDA NVidia], ALC662 Analog [ALC662 Analog]
[java] [SIP]
[java] [SIP] Aug 24 19:52:27.764 Using Speaker plughw:0,0
[java] [SIP err] speaker buffer size 10240., thresholdFrames 2555
[java] [SIP] Aug 24 19:52:27.851 Using Microphone plughw:0,0
[java] [SIP] Aug 24 19:52:27.911 Connected: Transmit media set to 1/16000/1, media payload is 103, Receiver media set to PCM/16000/2, media payload is 104
[java] [SIP err] speaker buffer size 10240., thresholdFrames 2555
[java] Error: entire set 8 passed through!
[java] Error: set size = 11
[java] Error: entire set 8 passed through!
[java] Error: set size = 9
[java] Error: entire set 8 passed through!
[java] Error: set size = 9
[java] Error: entire set 8 passed through!
[java] Error: set size = 12
[java] Error: entire set 8 passed through!
[java] Error: set size = 10
[java] Error: entire set 8 passed through!
[java] Error: set size = 10
[java] Error: entire set 8 passed through!
[java] Error: set size = 10
[java] Error: entire set 8 passed through!
[java] Error: set size = 10
[java] Error: entire set 8 passed through!
[java] Error: set size = 10
[java] Error: entire set 8 passed through!
[java] Error: set size = 10
[java] Error: entire set 8 passed through!
[java] Error: set size = 10
[java] Error: entire set 8 passed through!
[java] Error: set size = 9
[java] Error: entire set 8 passed through!
[java] Error: set size = 10
[java] Error: entire set 8 passed through!
[java] Error: set size = 9
[java] Error: entire set 8 passed through!
[java] Error: set size = 9
[java] Error: entire set 8 passed through!
[java] Error: set size = 11
[java] Error: entire set 8 passed through!
[java] Error: set size = 13
[java] Error: entire set 8 passed through!
[java] Error: set size = 9
[java] Error: entire set 8 passed through!
[java] Error: set size = 9
[java] Error: entire set 8 passed through!
[java] Error: set size = 9
[java] Error: entire set 8 passed through!
[java] Error: set size = 10
[java] Error: entire set 8 passed through!
[java] Error: set size = 9
[java] Error: entire set 8 passed through!
[java] Error: set size = 9
[java] Error: entire set 8 passed through!
[java] Error: set size = 9
[java] Error: entire set 8 passed through!
[java] Error: set size = 9
[java] Error: entire set 8 passed through!
[java] Error: set size = 9
[java] ColliisonUtil:getCollisionTriangles NOT SUPPORTED FOR TriangleStripArray
[java] ColliisonUtil:getCollisionTriangles NOT SUPPORTED FOR TriangleStripArray
[java] Error: entire set 8 passed through!
[java] Error: set size = 10
[java] ColliisonUtil:getCollisionTriangles NOT SUPPORTED FOR TriangleStripArray
[java] ColliisonUtil:getCollisionTriangles NOT SUPPORTED FOR TriangleStripArray
[java] ColliisonUtil:getCollisionTriangles NOT SUPPORTED FOR TriangleStripArray
[java] ColliisonUtil:getCollisionTriangles NOT SUPPORTED FOR TriangleStripArray
[java] ColliisonUtil:getCollisionTriangles NOT SUPPORTED FOR TriangleStripArray
[java] ColliisonUtil:getCollisionTriangles NOT SUPPORTED FOR TriangleStripArray
[java] ColliisonUtil:getCollisionTriangles NOT SUPPORTED FOR TriangleStripArray
[java] ColliisonUtil:getCollisionTriangles NOT SUPPORTED FOR TriangleStripArray
BUILD SUCCESSFUL
Total time: 3 minutes 25 seconds
I could post the server's output too, if it will help. I'm pretty sure the problem's in the client or my X server, though. I do see one message on the server side that's disturbing:
[java] INFO: getObject txn:TxnTrampoline[originalTxn:TransactionImpl[tid:3422]], oid:69, forUpdate:true failed due to deadlock: com.sleepycat.db.DeadlockException: DB_LOCK_DEADLOCK: Locker killed to resolve a deadlock: DB_LOCK_DEADLOCK: Locker killed to resolve a deadlock
But that appears when (just after) I disconnect the client.
Thanks,
Greg
|
|
|
|
|
|
|
|
|
Re: Performance issue on quad-core machine
Posted:
Aug 25, 2008 1:34 PM
in response to: lord_max
|
|
|
Hi Greg,
In xorg.conf the driver line should be
Driver "nvidia"
To be sure check the vendor string reported by glxinfo, it should say NVIDIA Corporation.
Rgds
Paul
|
|
|
|
|
|
|
|
|
Re: Performance issue on quad-core machine
Posted:
Aug 26, 2008 1:22 AM
in response to: paulby
|
|
|
Yes, I do have Driver "nvidia" in my xorg.conf, and I do see NVIDIA Corporation in the output of glxinfo.
For good measure, here's my xorg.conf (with comments removed):
Section "ServerLayout"
Identifier "Default Layout"
Screen "Default Screen" 0 0
InputDevice "Generic Keyboard"
InputDevice "Configured Mouse"
EndSection
Section "Files"
FontPath "/usr/share/fonts/X11/misc"
FontPath "/usr/X11R6/lib/X11/fonts/misc"
FontPath "/usr/share/fonts/X11/cyrillic"
FontPath "/usr/X11R6/lib/X11/fonts/cyrillic"
FontPath "/usr/share/fonts/X11/100dpi/:unscaled"
FontPath "/usr/X11R6/lib/X11/fonts/100dpi/:unscaled"
FontPath "/usr/share/fonts/X11/75dpi/:unscaled"
FontPath "/usr/X11R6/lib/X11/fonts/75dpi/:unscaled"
FontPath "/usr/share/fonts/X11/Type1"
FontPath "/usr/X11R6/lib/X11/fonts/Type1"
FontPath "/usr/share/fonts/X11/100dpi"
FontPath "/usr/X11R6/lib/X11/fonts/100dpi"
FontPath "/usr/share/fonts/X11/75dpi"
FontPath "/usr/X11R6/lib/X11/fonts/75dpi"
FontPath "/var/lib/defoma/x-ttcidfont-conf.d/dirs/TrueType"
EndSection
Section "Module"
Load "i2c"
Load "bitmap"
Load "ddc"
Load "extmod"
Load "freetype"
Load "glx"
Load "int10"
Load "vbe"
EndSection
Section "InputDevice"
Identifier "Generic Keyboard"
Driver "kbd"
Option "CoreKeyboard"
Option "XkbRules" "xorg"
Option "XkbModel" "pc104"
Option "XkbLayout" "us"
EndSection
Section "InputDevice"
Identifier "Configured Mouse"
Driver "mouse"
Option "CorePointer"
Option "Device" "/dev/input/mice"
Option "Protocol" "ImPS/2"
Option "Emulate3Buttons" "true"
EndSection
Section "Monitor"
Identifier "Generic Monitor"
HorizSync 28.0 - 51.0
VertRefresh 43.0 - 60.0
Option "DPMS"
EndSection
Section "Device"
Identifier "Generic Video Card"
Driver "nvidia"
Option "RenderAccel" "true"
EndSection
Section "Screen"
Identifier "Default Screen"
Device "Generic Video Card"
Monitor "Generic Monitor"
DefaultDepth 24
SubSection "Display"
Depth 1
Modes "1024x768" "800x600" "640x480"
EndSubSection
SubSection "Display"
Depth 4
Modes "1024x768" "800x600" "640x480"
EndSubSection
SubSection "Display"
Depth 8
Modes "1024x768" "800x600" "640x480"
EndSubSection
SubSection "Display"
Depth 15
Modes "1024x768" "800x600" "640x480"
EndSubSection
SubSection "Display"
Depth 16
Modes "1024x768" "800x600" "640x480"
EndSubSection
SubSection "Display"
Depth 24
Modes "1024x768" "800x600" "640x480"
EndSubSection
EndSection
And the output of glxinfo:
name of display:  .0
display: screen: 0
direct rendering: Yes
server glx vendor string: NVIDIA Corporation
server glx version string: 1.4
server glx extensions:
GLX_EXT_visual_info, GLX_EXT_visual_rating, GLX_SGIX_fbconfig,
GLX_SGIX_pbuffer, GLX_SGI_video_sync, GLX_SGI_swap_control,
GLX_EXT_texture_from_pixmap, GLX_ARB_multisample, GLX_NV_float_buffer,
GLX_ARB_fbconfig_float
client glx vendor string: NVIDIA Corporation
client glx version string: 1.4
client glx extensions:
GLX_ARB_get_proc_address, GLX_ARB_multisample, GLX_EXT_visual_info,
GLX_EXT_visual_rating, GLX_EXT_import_context, GLX_SGI_video_sync,
GLX_NV_swap_group, GLX_NV_video_out, GLX_SGIX_fbconfig, GLX_SGIX_pbuffer,
GLX_SGI_swap_control, GLX_NV_float_buffer, GLX_ARB_fbconfig_float,
GLX_EXT_fbconfig_packed_float, GLX_EXT_texture_from_pixmap,
GLX_EXT_framebuffer_sRGB, GLX_NV_present_video
GLX version: 1.3
GLX extensions:
GLX_EXT_visual_info, GLX_EXT_visual_rating, GLX_SGIX_fbconfig,
GLX_SGIX_pbuffer, GLX_SGI_video_sync, GLX_SGI_swap_control,
GLX_EXT_texture_from_pixmap, GLX_ARB_multisample, GLX_NV_float_buffer,
GLX_ARB_fbconfig_float, GLX_ARB_get_proc_address
OpenGL vendor string: NVIDIA Corporation
OpenGL renderer string: GeForce 6150SE nForce 430/PCI/SSE2/3DNOW!
OpenGL version string: 2.1.2 NVIDIA 173.14.09
OpenGL extensions:
GL_ARB_color_buffer_float, GL_ARB_depth_texture, GL_ARB_draw_buffers,
GL_ARB_fragment_program, GL_ARB_fragment_program_shadow,
GL_ARB_fragment_shader, GL_ARB_half_float_pixel, GL_ARB_imaging,
GL_ARB_multisample, GL_ARB_multitexture, GL_ARB_occlusion_query,
GL_ARB_pixel_buffer_object, GL_ARB_point_parameters, GL_ARB_point_sprite,
GL_ARB_shadow, GL_ARB_shader_objects, GL_ARB_shading_language_100,
GL_ARB_texture_border_clamp, GL_ARB_texture_compression,
GL_ARB_texture_cube_map, GL_ARB_texture_env_add,
GL_ARB_texture_env_combine, GL_ARB_texture_env_dot3, GL_ARB_texture_float,
GL_ARB_texture_mirrored_repeat, GL_ARB_texture_non_power_of_two,
GL_ARB_texture_rectangle, GL_ARB_transpose_matrix,
GL_ARB_vertex_buffer_object, GL_ARB_vertex_program, GL_ARB_vertex_shader,
GL_ARB_window_pos, GL_ATI_draw_buffers, GL_ATI_texture_float,
GL_ATI_texture_mirror_once, GL_S3_s3tc, GL_EXT_texture_env_add,
GL_EXT_abgr, GL_EXT_bgra, GL_EXT_blend_color,
GL_EXT_blend_equation_separate, GL_EXT_blend_func_separate,
GL_EXT_blend_minmax, GL_EXT_blend_subtract, GL_EXT_compiled_vertex_array,
GL_EXT_Cg_shader, GL_EXT_depth_bounds_test, GL_EXT_draw_range_elements,
GL_EXT_fog_coord, GL_EXT_framebuffer_blit, GL_EXT_framebuffer_multisample,
GL_EXT_framebuffer_object, GL_EXT_gpu_program_parameters,
GL_EXT_multi_draw_arrays, GL_EXT_packed_depth_stencil,
GL_EXT_packed_pixels, GL_EXT_pixel_buffer_object, GL_EXT_point_parameters,
GL_EXT_rescale_normal, GL_EXT_secondary_color,
GL_EXT_separate_specular_color, GL_EXT_shadow_funcs,
GL_EXT_stencil_two_side, GL_EXT_stencil_wrap, GL_EXT_texture3D,
GL_EXT_texture_compression_s3tc, GL_EXT_texture_cube_map,
GL_EXT_texture_edge_clamp, GL_EXT_texture_env_combine,
GL_EXT_texture_env_dot3, GL_EXT_texture_filter_anisotropic,
GL_EXT_texture_lod, GL_EXT_texture_lod_bias, GL_EXT_texture_mirror_clamp,
GL_EXT_texture_object, GL_EXT_texture_sRGB, GL_EXT_timer_query,
GL_EXT_vertex_array, GL_IBM_rasterpos_clip,
GL_IBM_texture_mirrored_repeat, GL_KTX_buffer_region, GL_NV_blend_square,
GL_NV_copy_depth_to_color, GL_NV_depth_clamp, GL_NV_fence,
GL_NV_float_buffer, GL_NV_fog_distance, GL_NV_fragment_program,
GL_NV_fragment_program_option, GL_NV_fragment_program2,
GL_NV_framebuffer_multisample_coverage, GL_NV_half_float,
GL_NV_light_max_exponent, GL_NV_multisample_filter_hint,
GL_NV_occlusion_query, GL_NV_packed_depth_stencil, GL_NV_pixel_data_range,
GL_NV_point_sprite, GL_NV_primitive_restart, GL_NV_register_combiners,
GL_NV_register_combiners2, GL_NV_texgen_reflection,
GL_NV_texture_compression_vtc, GL_NV_texture_env_combine4,
GL_NV_texture_expand_normal, GL_NV_texture_rectangle,
GL_NV_texture_shader, GL_NV_texture_shader2, GL_NV_texture_shader3,
GL_NV_vertex_array_range, GL_NV_vertex_array_range2, GL_NV_vertex_program,
GL_NV_vertex_program1_1, GL_NV_vertex_program2,
GL_NV_vertex_program2_option, GL_NV_vertex_program3,
GL_NVX_conditional_render, GL_SGIS_generate_mipmap, GL_SGIS_texture_lod,
GL_SGIX_depth_texture, GL_SGIX_shadow, GL_SUN_slice_accum
glu version: 1.3
glu extensions:
GLU_EXT_nurbs_tessellator, GLU_EXT_object_space_tess
(followed by a long and detailed and very boring table)
Thanks,
Greg
|
|
|
|
|
|
|
|
|
|
|
|
Re: Performance issue on quad-core machine
Posted:
Aug 26, 2008 8:09 AM
in response to: lord_max
|
|
Helpful |
|
|
HI Greg,
Hmm, very strange. The settings all look good to me, but your glxgears performance is still really low. I'll check my driver and extension settings when I get to the office just to be sure.
The X11 Render extension allows X11 2D rendering to use the GPU to accelerate 2D desktop apps, it won't effect Wonderland.
Can you see anything in the X11 logs ? There is usually some entries indicating the Nvidia driver is running, you may find some error conditions (although I doubt it).
Rgds
Paul
|
|
|
|
|
|
|
|
|
Re: Performance issue on quad-core machine
Posted:
Aug 26, 2008 9:54 PM
in response to: paulby
|
|
|
On Tue, 2008-08-26 at 08:09 -0700, wonderland@javadesktop.org wrote:
> HI Greg,
>
> Hmm, very strange. The settings all look good to me, but your glxgears
> performance is still really low. I'll check my driver and extension
> settings when I get to the office just to be sure.
>
> The X11 Render extension allows X11 2D rendering to use the GPU to
> accelerate 2D desktop apps, it won't effect Wonderland.
>
> Can you see anything in the X11 logs ? There is usually some entries
> indicating the Nvidia driver is running, you may find some error
> conditions (although I doubt it).
>
> Rgds
>
> Paul
Here's mine, I notice that the framebuffer isn't on?? At least that is
what I think it is.
Section "Module"
Load "dbe"
Load "extmod"
Load "fbdevhw" <--------
Load "glx"
Load "record"
Load "freetype"
Load "type1"
EndSection
That's all I have in mine, using an older GeForce FX-5200.
Plus mine is identified in this:
Section "Device"
Identifier "Videocard0"
Driver "nvidia"
VendorName "NVIDIA Corporation"
BoardName "GeForce FX 5200"
EndSection
Did you use the nVidia bin file from nVidia?? Let it write
the /e Ric
|
|
|
|
|
|
|
|
|
Re: Performance issue on quad-core machine
Posted:
Aug 27, 2008 12:49 AM
in response to: Ric Moore
|
|
|
Hi Ric. Thanks for the suggestions. I tried adding the fbdevhw module, and the Extensions section. I got these new lines in the X11 log:
(**) Extension "Composite" is disabled
(WW) Ignoring unrecognized extension "UseCompositeWrapper"
(II) LoadModule: "fbdevhw"
(II) Loading /usr/lib/xorg/modules/linux/libfbdevhw.so
(II) Module fbdevhw: vendor="X.Org Foundation"
compiled for 7.1.1, module version = 0.0.2
ABI class: X.Org Video Driver, version 1.0
No change to my glxgears FPS though.
As I mentioned, I'm using Debian, not Ubuntu. My kernel is indeed version 2.6.26, but I also have the nVidia driver version 173.14.09 mentioned in the article, so it should be supported. I did allow the installer to compile the kernel module, and to modify my xorg.conf. And I see it in the output of lsmod:
Module Size Used by
nvidia 7085636 24
agpgart 24980 1 nvidia
nfs 118860 0
ipv6 202500 16
nfsd 81576 13
lockd 53992 2 nfs,nfsd
nfs_acl 3008 2 nfs,nfsd
sunrpc 148992 12 nfs,nfsd,lockd,nfs_acl
exportfs 3968 1 nfsd
dm_mod 42052 0
usbhid 24448 0
snd_hda_intel 317588 1
snd_pcm_oss 33248 0
snd_mixer_oss 12608 1 snd_pcm_oss
snd_pcm 56836 2 snd_hda_intel,snd_pcm_oss
snd_timer 16776 1 snd_pcm
snd 40804 7 snd_hda_intel,snd_pcm_oss,snd_mixer_oss,snd_pcm,snd_timer
8250_pnp 9664 0
thermal 15580 0
ohci_hcd 18628 0
soundcore 5664 1 snd
ehci_hcd 27596 0
parport_pc 22692 0
8250 19780 1 8250_pnp
usbcore 105264 4 usbhid,ohci_hcd,ehci_hcd
snd_page_alloc 7496 2 snd_hda_intel,snd_pcm
parport 28140 1 parport_pc
processor 32132 1 thermal
thermal_sys 9320 2 thermal,processor
forcedeth 43920 0
button 6416 0
i2c_nforce2 5888 0
serial_core 15936 1 8250
i2c_core 17108 2 nvidia,i2c_nforce2
evdev 8192 3
And if I needed any additional confirmation, there's the fact that I get the NVIDIA splash screen when X-Windows starts. So I have to believe that the kernel module is talking to the hardware, and the X-Windows driver is talking to the module.
Here's the last bit of /var/log/nvidia-installer.log:
-> Searching for conflicting X files:
-> done.
-> Searching for conflicting OpenGL files:
-> done.
-> Installing 'NVIDIA Accelerated Graphics Driver for Linux-x86' (173.14.09):
executing: '/sbin/ldconfig'...
executing: '/sbin/depmod -aq'...
-> done.
-> Driver file installation is complete.
-> Running post-install sanity check:
-> done.
-> Post-install sanity check passed.
-> Shared memory test passed.
-> Running runtime sanity check:
-> done.
-> Runtime sanity check passed.
-> Would you like to run the nvidia-xconfig utility to automatically update you
r X configuration file so that the NVIDIA X driver will be used when you res
tart X? Any pre-existing X configuration file will be backed up. (Answer: Y
es)
-> Your X configuration file has been successfully updated. Installation of
the NVIDIA Accelerated Graphics Driver for Linux-x86 (version: 173.14.09) is
now complete.
Thanks,
Greg
|
|
|
|
|
|
|
|
|
Re: Performance issue on quad-core machine
Posted:
Aug 27, 2008 8:53 AM
in response to: lord_max
|
|
|
On Wed, 2008-08-27 at 00:49 -0700, wonderland@javadesktop.org wrote:
> Hi Ric. Thanks for the suggestions. I tried adding the fbdevhw
> module, and the Extensions section. I got these new lines in the X11
> log:
>
> (**) Extension "Composite" is disabled
> (WW) Ignoring unrecognized extension "UseCompositeWrapper"
>
> (II) LoadModule: "fbdevhw"
> (II) Loading /usr/lib/xorg/modules/linux/libfbdevhw.so
> (II) Module fbdevhw: vendor="X.Org Foundation"
> compiled for 7.1.1, module version = 0.0.2
> ABI class: X.Org Video Driver, version 1.0
>
> No change to my glxgears FPS though.
>
> As I mentioned, I'm using Debian, not Ubuntu. My kernel is indeed
> version 2.6.26, but I also have the nVidia driver version 173.14.09
> mentioned in the article, so it should be supported. I did
> allow the installer to compile the kernel module, and to modify my
> xorg.conf.
> Thanks,
> Greg
Greg, you may wish to see if you can install a pre-2.6.26 kernel, just
for grins. It shouldn't harm anything as you'll have just an extra
kernel to boot with. I'd like to know. In Fedora Users list there were
several other people running the new kernel experiencing similar
problems.
I'm running nVidia ver 169.12 and kernel ver 2.6.18-92.1.10.el5.
I've got a pile of windows open on 6 desktops and GLX gears is giving me
5586 frames in 5.0 seconds = 1117.164 FPS. As I mentioned this is using
an old GeForce FX-5200. Oh yeah, is this a video-card inserted into a
slot, or is this built-in to the motherboard?? The amount of ram set
aside for the built-ins is critical. But, if you didn't experience
problems before, and do now, I'd bet on the new kernel is messing things
up and the new patched nVidia driver not being completely 100%. YMMV,
Ric
|
|
|
|
|
|
|
|
|
Re: Performance issue on quad-core machine
Posted:
Sep 6, 2008 10:42 AM
in response to: Ric Moore
|
|
|
Thanks again. To answer your question: This is an nVidia chipset on the motherboard (ASUS M2N-MX SE Micro-ATX AM2). The NVIDIA X Server Settings dialog provides an answer as to the memory available to it:
Graphics Processor: GeForce 6150SE nForce 430
VBIOS Version: 05.61.32.25.00
Memory: 512 MB
Bus Type: Integrated
Bus ID: 0:13:0
IRQ: 11
X Screens: Screen 0
Display Devices: Sony CPD-100ES (CRT-0)
Not sure how I'd increase that 512 MB, if it's too small.
When you say "if you didn't experience problems before, and do now"... well, there was no 'before'. As I said, this is a dedicated Wonderland server box. I bought it for this purpose and installed Linux on it - it's never been used for anything else before. So I don't have any prior experience with which to compare.
I will try installing an older kernel and see if that helps. (I did try upgrading to the latest nVidia driver, 173.14.12.) Thanks for the suggestion.
|
|
|
|
|
|
|
|
|
Re: Performance issue on quad-core machine
Posted:
Sep 7, 2008 12:47 AM
in response to: lord_max
|
|
|
On Sat, 2008-09-06 at 10:42 -0700, wonderland@javadesktop.org wrote:
> Thanks again. To answer your question: This is an nVidia chipset on the motherboard (ASUS M2N-MX SE Micro-ATX AM2). The NVIDIA X Server Settings dialog provides an answer as to the memory available to it:
>
> Graphics Processor: GeForce 6150SE nForce 430
> VBIOS Version: 05.61.32.25.00
> Memory: 512 MB
> Bus Type: Integrated
> Bus ID: 0:13:0
> IRQ: 11
> X Screens: Screen 0
> Display Devices: Sony CPD-100ES (CRT-0)
>
> Not sure how I'd increase that 512 MB, if it's too small.
>
> When you say "if you didn't experience problems before, and do now"...
> well, there was no 'before'. As I said, this is a dedicated
> Wonderland server box. I bought it for this purpose and installed
> Linux on it - it's never been used for anything else before. So I
> don't have any prior experience with which to compare.
>
> I will try installing an older kernel and see if that helps. (I did
> try upgrading to the latest nVidia driver, 173.14.12.) Thanks for the
> suggestion.
I'm running 2.6.18-53 if that helps. I just installed CentOS (RedHat
based server) and it went straight on to update the kernel. Things went
crazy. It seems some kernel developers don't want "proprietary drivers"
latching straight into the kernel. Me, I don't care as I just want it to
work. <grins> This kernel works just fine for me, I hope this helps. You
might just give the video card 256 megs and let the system do it's thing
with the rest of the ram, if it's sharing it. That should be plenty for
Wonderland. Wonderland worked just fine with 128 megs. with my old
nVidia card.
Since it is to be a server-box, I'd just jam in a nVidia card and let
the Motherboard use all of it's memory for the system. I'm old school,
everything on a card and a modem external. If something goes wrong,
replace it for 20 bucks. I'm cheap. <grins> Ric
--
----------------------------------------------------
My father, Victor Moore (Vic) used to say:
"There are two Great Sins in the world...
..the Sin of Ignorance, and the Sin of Stupidity.
Only the former may be overcome." R.I.P. Dad.
Linux user# 44256 Sign up at: http://counter.li.org/
https://nuoar.dev.java.net/
Verizon Cell # 336-254-1339
|
|
|
|
|
|
|
|
|
Re: Performance issue on quad-core machine
Posted:
Sep 7, 2008 12:53 AM
in response to: lord_max
|
|
|
On Sat, 2008-09-06 at 10:42 -0700, wonderland@javadesktop.org wrote:
> Thanks again. To answer your question: This is an nVidia chipset on
> the motherboard (ASUS M2N-MX SE Micro-ATX AM2). The NVIDIA X Server
> Settings dialog provides an answer as to the memory available to it:
Oh yeah, quad core, I almost forgot. There's some issues, and I need to
research that... what kernel do you have installed?? And how much
memory? If >= 4 megs, there is a special added bit to your /boot/grub/
startup file. In RedHat-land it's grub.conf. In Ubuntu land
it's ...something not grub.conf, I forget. But, you have to tell the
kernel that you have 4 megs or greater. That'll make it weird, too. Ric
--
----------------------------------------------------
My father, Victor Moore (Vic) used to say:
"There are two Great Sins in the world...
..the Sin of Ignorance, and the Sin of Stupidity.
Only the former may be overcome." R.I.P. Dad.
Linux user# 44256 Sign up at: http://counter.li.org/
https://nuoar.dev.java.net/
Verizon Cell # 336-254-1339
|
|
|
|
|
|
|
|
|
Re: Performance issue on quad-core machine
Posted:
Sep 7, 2008 4:20 AM
in response to: Ric Moore
|
|
|
Hello Ric,
Thank you tremendously: You had the right answer, it was the kernel version. I had been running 2.6.26. When I tried 2.6.18.8 (aiming for something close to what you're using), suddenly I started getting much better performance in Wonderland. I had some issues with the NIC driver under that version, so next I tried 2.6.25.16, the latest version before 2.6.26. My NIC issue went away and the Wonderland performance remains good. glxgears now tells me I'm getting 312 FPS when run full-screen, rather than 295, but the effect on Wonderland is far more dramatic.
I is a happy camper now, and am looking forward to setting aside issues of machine configuration and beginning to really play with this thing. From my perspective, the rest of this is academic and you can skip it if you like. But hey, maybe it's possible to get even better performance, so I'll try to answer your questions again.
You asked about memory... here's the output of cat /proc/meminfo:
MemTotal: 1945132 kB
MemFree: 951348 kB
Buffers: 42156 kB
Cached: 566284 kB
SwapCached: 0 kB
Active: 628732 kB
Inactive: 246840 kB
HighTotal: 1048256 kB
HighFree: 357096 kB
LowTotal: 896876 kB
LowFree: 594252 kB
SwapTotal: 1028120 kB
SwapFree: 1028120 kB
Dirty: 188 kB
Writeback: 0 kB
AnonPages: 267180 kB
Mapped: 57492 kB
Slab: 55968 kB
SReclaimable: 45736 kB
SUnreclaim: 10232 kB
PageTables: 1516 kB
NFS_Unstable: 0 kB
Bounce: 0 kB
CommitLimit: 2000684 kB
Committed_AS: 801240 kB
VmallocTotal: 114680 kB
VmallocUsed: 48156 kB
VmallocChunk: 63476 kB
HugePages_Total: 0
HugePages_Free: 0
HugePages_Rsvd: 0
HugePages_Surp: 0
Hugepagesize: 4096 kB
I'd be happy to give more of that to the nVidia. It currently reports 512 MB.
When you say "If >= 4 megs, there is a special added bit to your /boot/grub/
startup file" - did you mean 4 gigs? I have two, as you can see from the output above. I'm using LILO rather than Grub, so any kernel options I need go in /etc/lilo.conf. Right now, I'm just using "nolapic".
Thanks again,
Greg
|
|
|
|
|
|
|
|
|
Re: Performance issue on quad-core machine
Posted:
Sep 7, 2008 10:14 AM
in response to: lord_max
|
|
|
On Sun, 2008-09-07 at 04:20 -0700, wonderland@javadesktop.org wrote:
> Hello Ric,
>
> Thank you tremendously: You had the right answer, it was the kernel
> version. I had been running 2.6.26. When I tried 2.6.18.8 (aiming
> for something close to what you're using), suddenly I started getting
> much better performance in Wonderland. I had some issues with
> the NIC driver under that version, so next I tried 2.6.25.16, the
> latest version before 2.6.26. My NIC issue went away and the
> Wonderland performance remains good. glxgears now tells me I'm
> getting 312 FPS when run full-screen, rather than 295, but the effect
> on Wonderland is far more dramatic.
>
> I is a happy camper now, and am looking forward to setting aside
> issues of machine configuration and beginning to really play with this
> thing. From my perspective, the rest of this is academic and you can
> skip it if you like. But hey, maybe it's possible to get even better
> performance, so I'll try to answer your questions again.
>
> You asked about memory... here's the output of cat /proc/meminfo:
Ah! You're going good, grasshopper! Next thing is to work on your glx
figures. You should be hitting around 1,000cps if not better. So, do you
have the "mesa" utilities and libs installed?? I'd grab all of them,
then re-run your nvidia install again. Here's the rub, with any Redhat
flavor of distro, I can open a root terminal and enter "init 3" to go to
runlevel 3 (command line, no X) in order to set up the nVidia install.
The nVidia install won't work correctly (AFAIK) unless X is shut down. I
haven't been able to do that in kubuntu. Another learning curve for me.
The idea is to install nVidia from level 3, let it make all the compiles
and re-write / install of /etc/X11/xorg.conf and then enter init 5 once
it's done. That puts you back into X startup and login. Sweet. With the
nVidia modules done correctly, and mesa libs installed, you should see a
great performance boost. I ran a Ge-Force GLX-5200 (worth about $20) and
got 1,200 CPS on 128megs of vram, once everything was running correctly.
This is without the fancy rotating cube for a desktop stuff on. Dink
with it somemore and you may have a pleasant surprise.
BTW, in your package manager be sure to lock your kernel version!!!
Otherwise, tomorrow morning it'll be updated to breakage-prone.
OH YEAH!! With your new old-kernel, you have to re-do your nvidia
package anyway, as the modules are kernel version dependent. Do that
first to see if that is where your performance is taking a hit. Shoot
for at least 1,000CPS. You may be good to go. Enjoy your Sunday! Ric
--
----------------------------------------------------
My father, Victor Moore (Vic) used to say:
"There are two Great Sins in the world...
..the Sin of Ignorance, and the Sin of Stupidity.
Only the former may be overcome." R.I.P. Dad.
Linux user# 44256 Sign up at: http://counter.li.org/
https://nuoar.dev.java.net/
Verizon Cell # 336-254-1339
|
|
|
|
|
|
|
|
|
Re: Performance issue on quad-core machine
Posted:
Sep 21, 2008 9:24 PM
in response to: Ric Moore
|
|
|
Hi Ric,
I should probably clarify that when I say "running glxgears full-screen", I mean by maximizing the window. On my desktop, that makes the window around 1015 x 715. The larger the window, the lower the FPS you get. In my case, I get around 236 FPS at 1015x715.
When I first start the program, it comes up with a size of 300x300. At that size, I get around 1,690 FPS. I'm happy with how it's performing, and have gone on to playing with the Wonderland code.
As far as the kernel business: Debian handles kernels a bit differently than Ubuntu. There are some pre-compiled kernels available, but I prefer to roll my own, directly from the source at kernel.org. Once I've configured it, I use "make-kpkg" to compile the kernel and modules, and generate a Debian package from them. I install that package, and I've got another kernel I can use. (The packages can be compiled on one machine and moved to another, which is sometimes handy.) Since the kernel package was created by me, it doesn't correspond with any standard Debian package, and will never be replaced with an upgraded version from a standard archive.
Thanks again for leading me to the answer.
Greg
|
|
|
|
|
|
|
|
|
Re: Performance issue on quad-core machine
Posted:
Sep 21, 2008 10:00 PM
in response to: lord_max
|
|
|
On Sun, 2008-09-21 at 21:24 -0700, wonderland@javadesktop.org wrote:
> Hi Ric,
>
> I should probably clarify that when I say "running glxgears
> full-screen", I mean by maximizing the window. On my desktop, that
> makes the window around 1015 x 715. The larger the window, the lower
> the FPS you get. In my case, I get around 236 FPS at 1015x715.
>
> When I first start the program, it comes up with a size of 300x300.
> At that size, I get around 1,690 FPS. I'm happy with how it's
> performing, and have gone on to playing with the Wonderland code.
>
> As far as the kernel business: Debian handles kernels a bit
> differently than Ubuntu. There are some pre-compiled kernels
> available, but I prefer to roll my own, directly from the source at
> kernel.org. Once I've configured it, I use "make-kpkg" to compile the
> kernel and modules, and generate a Debian package from them. I
> install that package, and I've got another kernel I can use. (The
> packages can be compiled on one machine and moved to another, which is
> sometimes handy.) Since the kernel package was created by me, it
> doesn't correspond with any standard Debian package, and will never be
> replaced with an upgraded version from a standard archive.
Heh, in the "GOOD OLE DAYS" that was how we got by. Now, with all the
extras, new stuff and complicated hoops to jump through, I don't even
try. Back when you really had a 386 kernel, if you had a 486, it was
quite a bit of difference to have a 486 kernel. Maybe I ought to give it
a try sometime! And, you're welcome although I don't know what I did to
help! Glad to meet a man that rolls-his-own. <grins> Ric
--
----------------------------------------------------
My father, Victor Moore (Vic) used to say:
"There are two Great Sins in the world...
..the Sin of Ignorance, and the Sin of Stupidity.
Only the former may be overcome." R.I.P. Dad.
Linux user# 44256 Sign up at: http://counter.li.org/
https://nuoar.dev.java.net/
Verizon Cell # 434-774-4987
|
|
|
|
|
|
|
|
|
Re: Performance issue on quad-core machine
Posted:
Aug 27, 2008 12:23 AM
in response to: paulby
|
|
|
Hi Paul, thanks for trying to help me out.
Since the Xorg.0.log is pretty long, I'll just paste the parts that seem interesting / relevant (blank lines indicate where I skipped ahead):
X Window System Version 7.1.1
Release Date: 12 May 2006
X Protocol Version 11, Revision 0, Release 7.1.1
Build Operating System: UNKNOWN
Current Operating System: Linux coin 2.6.26coin1 #1 SMP Sun Jul 27 20:38:36 PDT
2008 i686
Build Date: 29 May 2008
(==) ModulePath set to "/usr/lib/xorg/modules"
(WW) Open ACPI failed (/var/run/acpid.socket) (No such file or directory)
(II) No APM support in BIOS or kernel
(++) using VT number 7
(WW) xf86OpenConsole: setpgid failed: Operation not permitted
(WW) xf86OpenConsole: setsid failed: Operation not permitted
(--) PCI:*(0:13:0) nVidia Corporation unknown chipset (0x03d0) rev 162, Mem @ 0xde000000/24, 0xc0000000/28, 0xdd000000/24, BIOS @ 0xdffc0000/17
(II) LoadModule: "glx"
(II) Loading /usr/lib/xorg/modules/extensions/libglx.so
(II) Module glx: vendor="NVIDIA Corporation"
compiled for 4.0.2, module version = 1.0.0
Module class: X.Org Server Extension
(II) NVIDIA GLX Module 173.14.09 Thu Jun 5 00:07:40 PDT 2008
(II) Loading extension GLX
(II) LoadModule: "nvidia"
(II) Loading /usr/lib/xorg/modules/drivers/nvidia_drv.so
(II) Module nvidia: vendor="NVIDIA Corporation"
compiled for 4.0.2, module version = 1.0.0
Module class: X.Org Video Driver
(II) NVIDIA dlloader X Driver 173.14.09 Wed Jun 4 23:47:25 PDT 2008
(II) NVIDIA Unified Driver for all Supported NVIDIA GPUs
(II) Primary Device is: PCI 00:0d:0
(--) Assigning device section with no busID to primary device
(--) Chipset NVIDIA GPU found
(II) Loading sub module "wfb"
(II) LoadModule: "wfb"
(II) Loading /usr/lib/xorg/modules/libwfb.so
(II) Module wfb: vendor="NVIDIA Corporation"
compiled for 7.1.99.2, module version = 1.0.0
(II) Setting vga for screen 0.
(**) NVIDIA(0): Depth 24, (--) framebuffer bpp 32
(==) NVIDIA(0): RGB weight 888
(==) NVIDIA(0): Default visual is TrueColor
(==) NVIDIA(0): Using gamma correction (1.0, 1.0, 1.0)
(**) NVIDIA(0): Option "RenderAccel" "true"
(**) NVIDIA(0): Enabling RENDER acceleration
(II) NVIDIA(0): NVIDIA GPU GeForce 6150SE nForce 430 (C61) at PCI:0:13:0
(II) NVIDIA(0): (GPU-0)
(--) NVIDIA(0): Memory: 524288 kBytes
(--) NVIDIA(0): VideoBIOS: 05.61.32.25.00
(--) NVIDIA(0): Interlaced video modes are supported on this GPU
(--) NVIDIA(0): Connected display device(s) on GeForce 6150SE nForce 430 at
(--) NVIDIA(0): PCI:0:13:0:
(--) NVIDIA(0): Sony CPD-100ES (CRT-0)
(--) NVIDIA(0): Sony CPD-100ES (CRT-0): 350.0 MHz maximum pixel clock
(II) NVIDIA(0): Assigned Display Device: CRT-0
(II) NVIDIA(0): Validated modes:
(II) NVIDIA(0): "1024x768"
(II) NVIDIA(0): "800x600"
(II) NVIDIA(0): "640x480"
(II) NVIDIA(0): Virtual screen size determined to be 1024 x 768
(++) NVIDIA(0): DPI set to (100, 100); computed from -dpi X commandline option
(==) NVIDIA(0): Disabling 32-bit ARGB GLX visuals.
(--) Depth 24 pixmap format is 32 bpp
(II) NVIDIA(0): Initialized GPU GART.
(II) NVIDIA(0): Unable to connect to the ACPI daemon; the ACPI daemon may not
(II) NVIDIA(0): be running or the "AcpidSocketPath" X configuration option
(II) NVIDIA(0): may not be set correctly. When the ACPI daemon is
(II) NVIDIA(0): available, the NVIDIA X driver can use it to receive ACPI
(II) NVIDIA(0): events. For details, please see the "ConnectToAcpid" and
(II) NVIDIA(0): "AcpidSocketPath" X configuration options in Appendix B: X
(II) NVIDIA(0): Config Options in the README.
(II) NVIDIA(0): Setting mode "1024x768"
(II) Loading extension NV-GLX
(II) NVIDIA(0): NVIDIA 3D Acceleration Architecture Initialized
(II) NVIDIA(0): Using the NVIDIA 2D acceleration architecture
(==) NVIDIA(0): Backing store disabled
(==) NVIDIA(0): Silken mouse enabled
(**) Option "dpms"
(**) NVIDIA(0): DPMS enabled
(II) Loading extension NV-CONTROL
(==) RandR enabled
(II) Setting vga for screen 0.
(II) Initializing built-in extension MIT-SHM
(II) Initializing built-in extension XInputExtension
(II) Initializing built-in extension XTEST
(II) Initializing built-in extension XKEYBOARD
(II) Initializing built-in extension XC-APPGROUP
(II) Initializing built-in extension SECURITY
(II) Initializing built-in extension XINERAMA
(II) Initializing built-in extension XFIXES
(II) Initializing built-in extension XFree86-Bigfont
(II) Initializing built-in extension RENDER
(II) Initializing built-in extension RANDR
(II) Initializing built-in extension COMPOSITE
(II) Initializing built-in extension DAMAGE
(II) Initializing built-in extension XEVIE
(II) Initializing extension GLX
Thanks,
Greg
|
|
|
|
|
|
|
|
|
Re: Performance issue on quad-core machine
Posted:
Aug 27, 2008 1:29 AM
in response to: lord_max
|
|
|
OK, let me throw out another thread of conversation, and see if anyone tugs on it...
Does anyone here know how the NVIDIA is, or is not, affected by the use of LAPICs? This is my first SMP machine and I know very little about them. When I first installed Linux on this box, I got ugly panic messages on boot, like these:
ENABLING IO-APIC IRQs
..TIMER: vector=0x31 apic1=0 pin1=0 apic2=-1 pin2=-1
..MP-BIOS bug: 8254 timer not connected to IO-APIC
...trying to set up timer (IRQ0) through the 8259A ... failed.
...trying to set up timer as Virtual Wire IRQ... failed.
...trying to set up timer as ExtINT IRQ... failed  .
Kernel panic - not syncing: IO-APIC + timer doesn't work! Boot with apic=debug
and send a report. Then try booting with the 'noapic' option
I did, in fact, originally try "noapic" as the panic message suggested. While I could boot the system that way, I wasn't able even to get the bridge server to run - I'm guessing the JVM was expecting certain interrupts and never saw them. Booting with "nolapic" gives me a system that seems to work fine... but I'm left with the lingering doubt, that perhaps certain high-end features (like the hardware 3-D acceleration) won't work right with the LAPIC support disabled in the kernel.
I've tried reading up on APICs and LAPICs and it's rather beyond me.
Thanks,
Greg
|
|
|
|
|
java.net RSS
