UNSOLVED

decinos

updated

12 years ago

D

decinos

1 Rookie

•

2 Posts

0

68341

November 26th, 2012 23:00

PowerEdge Compatible GPU List

Hello,

I want to use Nvidia GPU on PowerEdge servers. I know R720 configurable with GPUs but are there any specific list of graphic cards with the server is compatible. I would like to use Tesla K10 and K20 graphic cards.

  • DELL-Chris H

    7 Practitioner

    •

    9682 Posts

    •

    48046 Points

    7051

    0

    Posted November 27th, 2012 05:00

    Decinos,

    Here is a link to the GPU Compatibility list. It will show the uses and the card models.

    Let me know if this helps.

    DELL-Chris H

    Social Media and Communities Professional

    Dell Technologies | Enterprise Support Services

    #IWork4Dell

    Did I answer your query? Please click on ‘Mark as Accepted Answer’. ‘Thumbs up’ the posts you like!

  • decinos

    1 Rookie

    •

    2 Posts

    7051

    0

    Posted December 3rd, 2012 03:00

    Chris,

    Sorry but this doesn't help.

    Because the page not include K10 and K20 graphic cards. Are theese cards tested?

  • theflash1932

    11 Legend

    •

    16255 Posts

    7051

    0

    Posted December 3rd, 2012 08:00

    Chris' reply should help, as it tells you which cards are supported/compatible.  The K10/K20 are not on the list, so they are not supported or officially compatible.  They may work, but it is up to you to try and to troubleshoot any issues, as Dell would likely not have tested them at all.

  • 7051

    0

    Posted February 25th, 2013 15:00

    I have successfully run the K10s in Linux on the R720s. There is a PCI error message that you can skip and you need to modify the BIOS so that the fan speeds run at max. We get good performance out of them on our seismic processing.

    Unfortunately, it doesn't work in Windows Server 2008 R2 SP1.

    Haven't tried the K20s in Linux yet, will do so soon. I suspect Linux will work.

    Trying to install the Windows Server 2008 R2 SP1 on the K20 results in a BSOD.

  • 7051

    0

    Posted April 11th, 2014 13:00

    I didn't run fans at max for the K40s we got recently. For a long time we thought it was driver issues with the card with the above below. Turns out the card would shut down because of overheating. So full speed it is. Centos 6.4 is working for me and I would imagine 6.5 would as well. Try the latest BIOS. sudo nvidia-smi Wed Feb 19 16:55:40 2014 +------------------------------------------------------+ | NVIDIA-SMI 331.20 Driver Version: 331.20 | |-------------------------------+----------------------+----------------------+ | GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. | |===============================+======================+======================| | 0 ERR! ERR! | ERR! ERR! | ERR! | |ERR! ERR! ERR! ERR! / ERR! | 54MiB / 11519MiB | ERR! ERR! | +-------------------------------+----------------------+----------------------+ | 1 ERR! ERR! | ERR! ERR! | ERR! | |ERR! ERR! ERR! ERR! / ERR! | 54MiB / 11519MiB | ERR! ERR! | +-------------------------------+----------------------+----------------------+ +-----------------------------------------------------------------------------+ | Compute processes: GPU Memory | | GPU PID Process name Usage | |=============================================================================| | 0 ERROR: GPU is lost | | 1 ERROR: GPU is lost | +-----------------------------------------------------------------------------+
  • 7051

    0

    Posted April 11th, 2014 13:00

    On linux we keep getting pci error 32993 in centOS_6.5_final,

    I tried disabling pci slots 6-7 and got it to boot a couple times but no luck now after several reboots,

    Also, do you really need to run the fans at max? 

  • 7051

    0

    Posted April 11th, 2014 14:00

    Hi,

    We dont have an error during posting but during the OS bootup, 

    Also, did u use both the 8-pin and 6-pin connectors? From my memory, and previous setups, given the expansion card has one 8pin we should only really require the one 8-pin. However, the poweredge 720 states to plug both the 8-pin and 6-pin in.

    thanks again for the replies,

  • 7051

    0

    Posted April 11th, 2014 14:00

    This what I recommend doing and it has worked for us.

    1. To ignore the warning, not stop during boot, F2->System BIOS->Miscellaneous Settings->F1/F2 Prompt on Error->Disabled

    Plug & Play Configuration Error:
    Prefetchable Memory Allocation
    Slot 4
    Bus#43/Dev#10/Func#0 Slot 4


    2. To speed up the system fans. F2->iDRAC Settings->Thermal-> (two items) “Thermal Base Algorithm -> Maximum performance (Performance optimized)”, and the “Fan Speed Offset -> High Fan Speed Offset”
  • 7051

    0

    Posted April 11th, 2014 14:00

    So you never had any issues with APEI 32993 during boot of centOS?

    Anything else you did to the bios? 

    Were updating the bios now and will do a fresh install of centOS, 

    Originally, we installed the K10 after CentOS was installed.

  • ConfGen

    1 Message

    1178

    0

    Posted June 6th, 2014 00:00

    Hi ACCELEWARE ,

    I do see exactly that error message. I have one K1 and one K2 plug in and in ESX only the K1 is visible.

    Of course, I can simply disable the warning but what is the reason for this error? How to fix it?

    Thanks