|
|
#1 | |
|
Registered User
Join Date: Jan 2004
Posts: 21
|
Hi,
I have a mainboard with an 8200 igp (namely the asus m3n78-vm) and I want to overclock it a bit, to improve VDPAU performance. My question is, what has more affect on VDPAU, to change the shader clock or the gpu clock? |
|
|
|
|
|
|
#2 | |
|
Registered User
Join Date: May 2007
Posts: 72
|
Quote:
I'm really interested in this topic. Basically I see 2 methods: -BIOS -nvidia-settings For nvidia-settings method: I already have some attempts (see my pots at: http://www.nvnews.net/vbulletin/showthread.php?t=134373) For BIOS option: I tried to do overclocking on Gigabyte GA-M85M-US2H as this mobo has "VGA core clock" overclocking in BIOS. Unfortunately it seems to have no any effect on VDPAU measured via qvdpautest tool. I already contected GIGABYTE about BIOS error. They don't believe me, told about not supporting Linux and asked me for switch to Windows. Excellent support, isn't ? I'm wonder - are You already able to modify 2D/3D clocks in Your ASUS mobo ? If so - how You do this ? For VDPAU overclocking: I believe video decoder is linked to 2D clock, while deniterlacers are linked to 3D clock (shaders). br |
|
|
|
|
|
|
#3 |
|
Registered User
Join Date: Feb 2009
Posts: 226
|
I am interested in how you can separately modify those clocks as well. In nvidia-settings, I only see GPU and RAM.
|
|
|
|
|
|
#4 | |
|
Registered User
Join Date: Nov 2008
Posts: 7
|
I think VDPAU and shaders are relatively independent: running CUDA and VDPAU tasks simultaneously does not affect performance of each other. So IMHO GPU clock affects VDPAU performance more than shader clock.
|
|
|
|
|
|
|
#5 |
|
Registered User
Join Date: Jan 2004
Posts: 21
|
|
|
|
|
|
|
#6 | |
|
*BANNED*
Join Date: Mar 2009
Posts: 6
|
I can modify those clocks in the mainboard bios
![]() |
|
|
|
|
|
|
#7 | |
|
Registered User
Join Date: May 2007
Posts: 72
|
Quote:
http://hftom.free.fr/qvdpautest-0.3.tar.gz for tool download and http://www.nvnews.net/vbulletin/showthread.php?t=133465 for testing results. BTW: I had this mobo sometime ago. Was forced to sell it as it had problems with suspend2ram under Linux. Are You using suspend2ram ?. |
|
|
|
|
|
|
#8 | |
|
Registered User
Join Date: Jan 2004
Posts: 21
|
Quote:
Yes, got it working a few days ago. I had to disable message signaled interrupts (MSI), otherwise the harddrive wouldn't wake up on resume. I also disabled in-kernel agpgart to use NvAGP, don't know if that step was needed though. |
|
|
|
|
|
|
#9 | |
|
Registered User
Join Date: May 2007
Posts: 72
|
Quote:
Mine was - but only when no IR USB was attached on suspend. With attached MCE IR USB dongle - only 1 S2Ram cycle was possible. Every consecutive cycle ended with immediate resume. Disconnecting USB IR gives reliable S2Ram. I tried 3 different kernels (.25, .26., 27) and every BIOS release. As I'm using MCE USB to power-on from suspend - this issue marked this mobo as nonaceptable for me.... br |
|
|
|
|
|
|
#10 | |
|
Registered User
Join Date: Jan 2004
Posts: 21
|
I didn't test that much, but it survived several cycles with a Webcam plugged in. I don't use an IR USB device, I use the port from my dvb-t card. Don't know, if it is able to wake up the pc, I guess not...
I did some clocking: http://pastebin.com/f1cac7547 As you can see, the results are exactly as piotro predicted: Quote:
|
|
|
|
|
![]() |
| Thread Tools | |
|
|