|11-27-08, 07:21 PM||#1|
Join Date: Feb 2006
Onboard 8300 Performance with 178.82
Ok after trying every possible tweak (and obviously not the right one) regarding the nvidia driver and with different driver versions it is quite obvious to me that something is VERY wrong.
Loading the following picture as an example:
in Fedora 10 (gnome) with an Onboard 8300 using the 178.82 driver and the regular tweaks load slowly and occasionally lock xorg at 100% CPU load (one core) for a while. The lock is bad because it locks the entire xorg server so I have to wait for it to unlock before I can continue to use the computer which is unbelievably annoying.
If I use the vesa driver in the xorg server it loads as if it was nothing... no apparent pausing nor locking. Even when I switch workspaces I don't see the repainting happening as I do with the nvidia proprietary driver... its literally as bad as Windows 95 refresh back in the bad old day (yes it is that horrid.)
The above locking is even more apparent when switching workspaces and hitting the Firefox window space. It locks for a couple of seconds.
It becomes even more pronounced after I start k3b to a point of it being ridicules... I haven't been single tasking since 1984 on the old C64.
The graphic card can handle it because the vesa mode has definitely proved that to me... I did not have this issue on my 6150 older machine (that fried for some unknown reason) at all BUT I used an older driver for it.
I have 4GB ram and the video is set to 512MB... 4 core 9950... which again has no bearing because the vesa driver is definitely two to three magnitudes faster (ok ok at the minimum 2 magnitudes faster.)
Sadly because the motherboard that I have does not have a good VGA connector, it has a lot of static due to the VGA cords going more or less "naked" from the middle of the board to the connector in the case (Yey Asus for that magnificent design decision!) so I must use the DVI port to get any quality signal to my monitor THUS I think I must use the nvidia proprietary driver.
If anyone has a way to get that DVI port working with the vesa driver it would be VERY much appreciated!
Well if the nvidia driver got fixed it would also be VERY much appreciated BUT I believe that is harder than making the vesa DVI port aware.
In times like these I miss my Mach64, Matrox G400 and GeForce 3 no nonsense cards.
|12-24-08, 03:49 AM||#2|
Join Date: Dec 2008
Re: Onboard 8300 Performance with 178.82
Can't help you with your VESA / DVI problem, but I'm having the same issues with X here. Very slow repaint and freezes with your example pic.
I just upgraded my motherboard from an ASUS M2N-MX (GeForce 6100/nForce 430) to an ASUS M3N78-EM (GeForce 8300). I had no problems like this with the old board.
Running Ubuntu Intrepid with KDE 4. At first I kept the 177.X stable driver. Than I upgraded to the latest 180.17 BETA driver. Result: Still slow repainting and freezes. But with the BETA driver and the patched mplayer VDPAU works
Any pointers on how I can improve my desktop experience back to how it was?
|12-24-08, 10:53 AM||#3|
Join Date: Apr 2008
Re: Onboard 8300 Performance with 178.82
I guess it's a problem with the driver and KDE(4), because there's no such freeze on my system: Asus M3N78-EM, Phenom 9600, 8 GiB of RAM (512 MiB for the on-board GeForce 8300), running linux 22.214.171.124, NVIDIA drivers 177.82, Openbox.
I simply followed the optimization instructions in one of the pinned threads. Specifically, nvidia-settings -a :0/InitialPixmapPlacement=2 -a :0/GlyphCache=1 and Option "PixmapCacheSize" "10485760", Option "AllowSHMPixmaps" "0", Option "Composite" "Disable" in my xorg.conf (that cache size corresponds to 1280x1024x8 pixels, it's probably overkill).
Edit: actually it does take a second (with one CPU core at 100%) to redraw upon switching workspaces, but only after the Firefox workspace has been inactive for more than 15 seconds. To be clear, if I switch from Firefox's workspace to another one, wait 10 seconds, and come back to Firefox's workspace, the switch is instantaneous; if I wait 15 seconds or more, the switch takes about a second.
I observe the same behavior with my discrete 8600GT, although the lock is shorter, probably due to either the chip's higher performance, or the presence of actual video RAM.
|Thread||Thread Starter||Forum||Replies||Last Post|
|Video: Xyratex ' The Next Gen in Storage Performance and Efficiency||News||Latest Tech And Game Headlines||0||06-17-12 10:50 AM|
|Boost Your Performance Goals 10x ' This Week on inside* Publications||News||Latest Tech And Game Headlines||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|
|RAID question?||MUYA||CPUs, Motherboards And Memory||7||09-20-02 02:37 PM|