Results 1 to 6 of 6
  1. #1
    ROG Member Array
    Join Date
    Oct 2018
    Reputation
    10
    Posts
    6

    Issue with 2080ti card(s) causing 'code 92 TEST NVRAM' if used on slot PCIE16X_1

    Having a strange issue with only 2080ti cards on my Zenith Extreme. I've updated to the provided beta 1501 bios from elmor via overclockers.net forum for resolution of the nvlink '0E Vga Bios Load' error (which is now resolved), however this NVRAM Test error occured even before the beta bios update with 1402.

    Issue only occurs if a 2080ti is inserted into slot PCIE16X_1. I've tried (2) seperate 2080ti's on this port with same issue, both in single gpu and multipgpu/nvlink mode. Issue does not occur if a 980ti or TitanXp/1080ti is inserted into the same PCIE16X_1 slot and system will run fine without errors. Again, issue only occurs if a 2080ti card is inserted in the PCIE16x_1 slot. It does not matter if it is single gpu or multgpu with another 2080 in another slot, issues will present. IF 2080ti card(s) are moved to any other slot or slot configuration that does not use PCIE16x_1 port, system functions normally without issue and no errors. This, however, limits one card to 8x if used in multigpu nvlink.

    symptoms: mobo is continuously throwing a 92 TEST NVRAM code, with delayed posts, system halts after post, issues after windows loads (freezing, screen goes black, system reboots, windows runs at like 1fps when moving mouse.

    I have tried everything, clearing CMOS, default settings, even stripping components down to the bare minimum, only KB/M plugged into USB, nothing else except ram and NVMe drive plugged into mobo. Same outcome.

    Any suggestions or assistance at all would be greatly appreciated. This issue was present in 1402, and still in the 1501 beta bios released today.

    Also for reference are my system specs:

    Asus Zenith Extreme x399 (1501 beta bios provided by Elmor, previoulsy 1402 with same NVRAM issue)
    AMD Threadripper 2950x @ stock speed/precision boost
    Artic LiquidFreezer 360mm AIO CPU cooler
    EVGA SuperNOVA G3 1200w PSU
    Corsair Vengenance RGB 32GB DDR4 (8GBx4) 3600MHZ (clocked to 3200mhz) with modules in DIMM C1, D1, A1, B1 slots.
    Samsung 960 Pro 512GB NVMe - OS partition
    2xSamsung 950 SSD 1TB in RAID0 - Game partition
    2xWestern Digital 1TB HDD - Data partitions
    2xNvidia Geforce 2080ti FE RTX GPU's with NVLINK on a LG 4k HDR 65' TV monitor (via hdmi)
    Deepcool Quadstellar ATX case
    Corsair gaming mouse via USB
    Bloody mechanical Laser-switch USB keyboard
    Windows 10 Pro x64 with latest updates, drivers, etc.

  2. #2
    ROG Enthusiast Array
    Join Date
    May 2018
    Reputation
    10
    Posts
    31

    this error occurs due to sag or defects in the slot that may not occur with thicker cards, or grease / foreign matter in the slot or oxidised pins.

    is this slot running a riser cable?
    Last edited by Squall Leonhart; 10-12-2018 at 03:24 PM.

  3. #3
    ROG Member Array
    Join Date
    Oct 2018
    Reputation
    10
    Posts
    6

    Yes they use the pec 300 extension cable for pci-x slot so that the graphics cards can reach the motherboard from the quadrant that they're installed on.

    But that's only if you want to use that upper quadrant, else they can be inserted on the motherboard directly.

    I've tried eliminating the extension cable and just plugging the card(s) directly to the board. Same results. what you say may be accurate is I originally suspected it could have been a bad GPU and then I Ruled that out, then I suspected it was the extension cable, then I Ruled that out. So now it's either down to the slot is bad as you say, or another bios bug entirely which I'm not sure about.

    I'll check for any foreign material or corrosion on the slot pins. The motherboard currently is housed in a deepcool quadstellar case, that was Prior installed in a Cosmos 2 full-tower ATX case. At one point it did have the Titan Xps directly inserted on the motherboard in SLI, but since moving to the quad Stellar it has been using the riser cables to the PCI ports. The board was bought on the same day that the 1950x launched last year.

  4. #4
    ROG Guru: Yellow Belt Array
    Join Date
    Mar 2017
    Reputation
    13
    Posts
    132

    One thing you could try, is reflashing UEFI by using BIOS Flashback to version 0305 (boot computer so EC updates), and then BIOS flashback to 1501.

  5. #5
    ROG Member Array
    Join Date
    Oct 2018
    Reputation
    10
    Posts
    6

    Ended up being the PCIX Riser cables ( PEC 300 adapters ) included with the Deepcool Quadstellar case. They are apparently incompatible with the 2080ti series.

    Fortunately can run the cards in the case directly on the board itself ( though kind of defeats the purpose ), and once inserted, all slots are functioning fine in nvlink now


    Thanks for the help all!

  6. #6
    ROG Enthusiast Array
    Join Date
    May 2018
    Reputation
    10
    Posts
    31

    Quote Originally Posted by captaintrips View Post
    Ended up being the PCIX Riser cables ( PEC 300 adapters ) included with the Deepcool Quadstellar case. They are apparently incompatible with the 2080ti series.

    Fortunately can run the cards in the case directly on the board itself ( though kind of defeats the purpose ), and once inserted, all slots are functioning fine in nvlink now


    Thanks for the help all!
    the included cable looks like one of the cheaper shielded types that uses copper tape as shielding,

    for instance, the LiHeat
    Click image for larger version. 

Name:	98SR8Oyl.jpg 
Views:	16 
Size:	17.9 KB 
ID:	76273

    Twin Axiel or polymer shielding would be better to use on the 2080ti

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •