long lags when clicking on menu, fedora 12

[stevesz]

I have over 10 sec lag when clicking on menu icon, config icons etc.
I have two 8800GT with a SLI bridge.

X will hang after a while. But I can still ssh in remotely.

Odd things has showed up, like 0 CPU load but all three load averages showed 1.00. All four cores were at 0.00. - Disappeared after reboot.

Initially it loaded the nouveau driver but I renamed initd and created a new one. Added a line to point to extensions/nvidia rather than extensions. Then moved the extensions/libglx.so to /root to ensure it would not be loaded.

# rpm -qa | grep nvidia
akmod-nvidia-190.42-1.fc12.4.x86_64
nvidia-xconfig-1.0-1.fc12.x86_64
nvidia-settings-1.0-3.2.fc12.x86_64
xorg-x11-drv-nvidia-libs-190.42-5.fc12.x86_64
kmod-nvidia-2.6.31.6-145.fc12.x86_64-190.42-1.fc12.6.x86_64
xorg-x11-drv-nvidia-190.42-5.fc12.x86_64

akmod is v190.42-1 while nvidia-libs is -5 maybe that is the problem.

# Xorg configuration created by livna-config-display

Section "Files"
ModulePath "/usr/lib64/xorg/modules/extensions/nvidia"
ModulePath "/usr/lib64/xorg/modules/drivers"
ModulePath "/usr/lib64/xorg/modules"
EndSection

Section "ServerFlags"
Option "AIGLX" "on"
EndSection

Section "Device"
Identifier "Videocard0"
Driver "nvidia"
Option "AddARGBGLXVisuals" "True"
EndSection

Section "Extensions"
Option "Composite" "Enable"
EndSection

I ran:
akmods --kernels 2.6.31.6-145.fc12.x86_64
service nvidia start
chkconfig --level 5 nvidia on

All were OK.

Not sure how to proceed.

---------- Post added at 02:16 PM CST ---------- Previous post was at 11:32 AM CST ----------

I just let it sit and the screen locked up, showing a "carpet" of all colors. Plus these warnings showed which might just be a symptom.

(WW) Dec 04 12:45:37 NVIDIA(2): WAIT (2, 6, 0x8000, 0x000042a8, 0x000042b8)
(WW) Dec 04 12:45:44 NVIDIA(2): WAIT (1, 6, 0x8000, 0x000042a8, 0x000042b8)
(WW) Dec 04 12:45:47 NVIDIA(2): WAIT (2, 6, 0x8000, 0x00005e3c, 0x00005e4c)
[mi] EQ overflowing. The server is probably stuck in an infinite loop.

Backtrace:
0: /usr/bin/X (xorg_backtrace+0x28) [0x49e8d8]
1: /usr/bin/X (mieqEnqueue+0x1f4) [0x49e2a4]
2: /usr/bin/X (xf86PostMotionEventP+0xce) [0x478f0e]
3: /usr/lib64/xorg/modules/input/evdev_drv.so (0x7f91aa014000+0x50bf) [0x7f91aa0190bf]
4: /usr/bin/X (0x400000+0x6be17) [0x46be17]
5: /usr/bin/X (0x400000+0x116b13) [0x516b13]
6: /lib64/libpthread.so.0 (0x35ad400000+0xefa0) [0x35ad40efa0]
7: /usr/lib64/xorg/modules/drivers/nvidia_drv.so (0x7f91ca90b000+0x5da1d) [0x7f91ca968a1d]
8: /usr/lib64/xorg/modules/drivers/nvidia_drv.so (_nv001018X+0x28b) [0x7f91ca96939b]
9: /usr/lib64/xorg/modules/drivers/nvidia_drv.so (0x7f91ca90b000+0xa5640) [0x7f91ca9b0640]
10: /usr/lib64/xorg/modules/drivers/nvidia_drv.so (0x7f91ca90b000+0x349bae) [0x7f91cac54bae]
11: /usr/bin/X (0x400000+0xd0c5a) [0x4d0c5a]
12: /usr/bin/X (0x400000+0x70a95) [0x470a95]
13: /usr/bin/X (0x400000+0xb1d2b) [0x4b1d2b]
14: /usr/bin/X (ValidateGC+0x24) [0x43f534]
15: /usr/bin/X (0x400000+0x2a5b8) [0x42a5b8]
16: /usr/bin/X (0x400000+0xb6071) [0x4b6071]
17: /usr/bin/X (0x400000+0x2c69c) [0x42c69c]
18: /usr/bin/X (0x400000+0x21cfa) [0x421cfa]
19: /lib64/libc.so.6 (__libc_start_main+0xfd) [0x35acc1eb1d]
20: /usr/bin/X (0x400000+0x218a9) [0x4218a9]
(WW) Dec 04 12:45:54 NVIDIA(2): WAIT (1, 6, 0x8000, 0x00005e3c, 0x00005e4c)
(WW) Dec 04 12:45:58 NVIDIA(2): WAIT (2, 6, 0x8000, 0x0000edcc, 0x0000eddc)
(WW) Dec 04 12:46:05 NVIDIA(2): WAIT (1, 6, 0x8000, 0x0000edcc, 0x0000eddc)
(WW) Dec 04 12:51:09 NVIDIA(0): WAIT (2, 6, 0x8000, 0x00005550, 0x00005e60)

[marko]

I have this exact problem too. Just tonight I switched from nouveau to the rpmfusion driver for my 8800GT Nvidia card because I wanted suspend to ram to work (nouveau comes out of suspend but the video card fan goes to 100% speed permanently). I suppose it might fix this to go to the fully closed Nvidia "run" driver from nvidia but I want to avoid that if possible. I at least want either the rpm fusion or the closed driver because I want to use CUDA for programming.

Mark

UPDATE: I worked around this by going back to the nouveau driver.... it works ok but no CUDA and no suspend to RAM. I'll decide some time to go to the blob driver (assuming that doesn't also have this slow menu issue that the rpmfusion version has)

[lmcogs]

I had this problem but seems to be ok now but I don't remember how it was worked around other that it was something to do with a fix in xorg. I think it was either of these sites
http://www.nvnews.net/vbulletin/showthread.php?t=141514
https://bugzilla.redhat.com/show_bug.cgi?id=533620
I am using f12, nvidia driver and kde