Tyan S7025 graphics issue

Hi, I’ve just built a tesla machine using the tyan s7025 motherboard. I’ve installed ubuntu 9.04 but I can only get 800x600 resolution from the onboard graphics (which is an ASPEED AST2050 chipset). Does anyong have any ideas how to get this onboard graphics working properly in ubuntu??
I’ve updated the drivers to the latest ones which apparently support redhat and susi but there is no mention of ubuntu.

Cheers

If I’m not mistaking this is an CUDA-forum. You’ll probably get better support in a dedicated ubuntu/Linux forum. Anyway it’s not very easy to help you if there are so little information around. Could you please post your /etc/X11/xorg.conf file? There are probably no other resolutions set. Make sure that your “Display” subsection of your “Screen” section has a line that looks like this:

Modes	  "1280x1024" "1024x768" "800x600" "640x480"

You’ll have to enter your screen resolution there. E.g. for a screen with 1600 by 1200 pixel it should look like this:

Modes	  "1600x1200" "1280x1024" "1024x768" "800x600" "640x480"

–Teewurst

Well yes this is true, I posted here as this is a specialist motherboard recommended by NVIDIA for building a cuda PSC. I also posted in the ubuntu forum and got no reply.

Anyway I have resolved this issue and for others facing the same problem the solution is to download the latest drivers for the onboard graphics, then come the fun bit, you have to install a 32-bit version of windows, I used XP. from windows you can use the flash untility to update the VBIOS on the graphics chipset. Then back re-install ubuntu and hay presto you can now get high resolution on your display. No openGL though so non of the nice pretty cuda demos will work.

Next problem… despite installing the NVIDIA 190 driver successfully the tela cards are not detected!

Apparently I need to disable the motherboard VGA graphics from the BIOS settings, then install the NVIDIA driver (using a remote connection as there is now no display). Having done this I should be able to enable the graphics again and be ready to go. Fingers crossed.

Everything there is correct except the last bit, I fear. You probably won’t be able to use the onboard graphics with the Telsa cards at all.

I don’t know for sure but I think Linux does support more than one Graphics card at a time. You’ll sure need both drivers installed and then, if you choose your onboard-chip for X11, you should still be able to access the other card via CUDA. But I have no experience with such a setup or any other multi card setup. Maybe the guys from [url=“http://www.linuxquestions.org/”]http://www.linuxquestions.org/[/url] can help you.

–Teewurst

I am using the same motherboard with bios v1.03. The on-board VGA works just fine with Tesla C1060 and GeForce cards as well - using the driver v190.53. I didn’t have to flash the VBIOS either but as far as I remember the documentation mentions something about this being necessary in some cases.

When using GeForce cards I had to fiddle a bit with the xorg.conf, but is definitely possible to make it work. To force the OS to use the ASPEED this is my Device section:

Section "Device"

		Identifier  "Videocard0"

		Driver	  "ast"

		BusID	   "01@00:00:0"

EndSection

and in the Screen section I manually select the device, monitor and also the mode defined in the Monitor section:

Section "Monitor"

	Identifier	 "Monitor0"

	HorizSync	   30.0 - 81.0

	VertRefresh	 56.0 - 76.0

	ModeLine "1280x1024@60" 135.0 1280 1328 1440 1688 1024 1025 1028 1066 +hsync +vsync	

EndSection

Section "Screen"

		Identifier  "Screen0"

		Device	  "Videocard0"

		Monitor	 "Monitor0"

		DefaultDepth	 24

		SubSection "Display"

			Viewport	0 0

			Depth	   24

			Modes	   "1280x1024@60"

		EndSubSection

EndSection

Note that the ModeLine is specific to the monitor so you need to look up the correct values for your monitor.

When using only Tesla cards, as the on-board one is the only VGA adapter X does not get confused and everything works without fiddling with manual xorg.conf tweaking.

Cheers, it’s good to know it can be done. Did you have to disable the onborad graphics to install the NVIDIA 190 driver?? and then turn the onborad graphics back on??

It seems I’ve installed the 190 driver but the tesla card is not being detected.

So, how did you fix this?

I am having the same problem now.