PCI-E x16 @ x2 bug

Discussion in 'Videocards - NVIDIA GeForce' started by akbaar, Jun 2, 2020.

  1. akbaar

    akbaar Master Guru

    Messages:
    426
    Likes Received:
    55
    GPU:
    ASUS TUFF 3080 12Gb
    [​IMG]

    this problem popped up, from no where,,
    I tried Reseating the cpu, reseating the gpu , cleaned the contacts with legit cleaners from good source,
    tried reflashing motheboard BIOS with diffrent versions, reset to defaults, flashed the the gpu with different bioses too, changed cables,

    any ideas? you think that ME firmware is the culprit ? what is the problem u might think ?
    do you think the motherboard is shortcutted somehow? or the gpu is failing ? although am playing games good but i know it is not going over 60fps and am getting slight delays somtimes,
     
  2. The Goose

    The Goose Ancient Guru

    Messages:
    3,057
    Likes Received:
    375
    GPU:
    MSI Rtx3080 SuprimX
    With out knowing your specs I`d say youve already used up you pci-e lanes on nvme/ssd
     
  3. SpajdrEX

    SpajdrEX Ancient Guru

    Messages:
    3,399
    Likes Received:
    1,653
    GPU:
    Gainward RTX 4070
    Hard to say, it could be failing gfx card or PCIE slot is not working correctly anymore.
    You already tried to reset the GPU card, so now it would be good to try your card in some other motherboard if possible.
     
    akbaar likes this.
  4. akbaar

    akbaar Master Guru

    Messages:
    426
    Likes Received:
    55
    GPU:
    ASUS TUFF 3080 12Gb
    yeah i didnt have that problem before, i have mini ITX ASUS Maximus impactVIII with one PCI-e x16 and one NVME intel 750 i even disconnect the NVME and disabled it in the bios still problem exists.

    will try my best to try out the gpu on other motherboard, but for now i dont have spare motherboard to try on so it will take time to find spare motherboard outside the covid.

    hmm what else is there to try out to do with motherboard or gpu? should i downgrade ME firmware? because it was ok on my current ME firmware, hmm
     

  5. tsunami231

    tsunami231 Ancient Guru

    Messages:
    14,725
    Likes Received:
    1,854
    GPU:
    EVGA 1070Ti Black
    why would me firmware have any thing to do with this?
     
  6. akbaar

    akbaar Master Guru

    Messages:
    426
    Likes Received:
    55
    GPU:
    ASUS TUFF 3080 12Gb
    not sure, just guessing because i updated my ME firmware fairly recent .. with this version
    http://www.station-drivers.com/inde...tory&Itemid=353&func=fileinfo&id=4359&lang=en
    Intel Management Engine (ME) Firmware Version 11.8.77.3664 (S&H)(1.5Mo)

    i also have USB3.1 and had it "enabled always" in the BIOS, and used some SBM board to feed it power, do you think that it may have shortcuted the motherboard.. because in the BIOS discription, there was some warning having the port always on ... and it may be dangerous and harm the motherboard.


     
  7. akbaar

    akbaar Master Guru

    Messages:
    426
    Likes Received:
    55
    GPU:
    ASUS TUFF 3080 12Gb
    am not sure yet,
    [​IMG]
    [​IMG]

    i still cant pin point the problem, changed the power cable, changed wall socket, removed little pressure from cpu by the watercooler, cleaned the pci-e slot,
    do the iGPU got something to do with it ? hmmm

    edit: you think there is powersaving feature i dont know about maybe is doing this? how is that it states that my mobo supports 16x , but its only receiving 2x
     
  8. akbaar

    akbaar Master Guru

    Messages:
    426
    Likes Received:
    55
    GPU:
    ASUS TUFF 3080 12Gb
    i tried cleaning the cpu bottom part of it, still same issue, hmmm
     
  9. 386SX

    386SX Ancient Guru

    Messages:
    2,075
    Likes Received:
    2,238
    GPU:
    AMD Vega64 RedDevil
    ME update could have done this. Do you have the old version to revert to?
    Did you get any warning during flash itself? What do you mean by BIOS (where you got the warning)? Don't tell me you flashed ME manually and ignored a warning during flash ... only flash if there is NO warning at all.

    Good luck! ;)
     
    akbaar likes this.
  10. akbaar

    akbaar Master Guru

    Messages:
    426
    Likes Received:
    55
    GPU:
    ASUS TUFF 3080 12Gb
    yes i have old ME .
    yes i did manually but am sure there was no warning , i usually do it on regular basis with no problems.
    hmmm if ME is the problem i will need to do remove bios chip and reformat it manually using usb device.

    Edit: do you know why ME would have been the problem?
     
    Last edited: Jun 6, 2020

  11. 386SX

    386SX Ancient Guru

    Messages:
    2,075
    Likes Received:
    2,238
    GPU:
    AMD Vega64 RedDevil
    Regarding to why I suspected ME being the root cause:
    ME is the central management component which manages all your devices on your mainboard like wireless, AMT, chipset, thermal guards, Watchdog(-s), etc.

    Did you flash an official BIOS (then I wouldn't suspect ME being the culprit) or did you flash only the ME oprom as 1.5 or 5MB bin file?
     
  12. Astyanax

    Astyanax Ancient Guru

    Messages:
    17,011
    Likes Received:
    7,353
    GPU:
    GTX 1080ti
    ME does not manage anything to do with the cpu's PCIE controller, it only has a hand in components integrated on the mainboard.
     
  13. akbaar

    akbaar Master Guru

    Messages:
    426
    Likes Received:
    55
    GPU:
    ASUS TUFF 3080 12Gb
    ok... got it... somebody said bent motherboard pins for cpu can cause the cpu to hold off on the pci-e lanes...

    what am thinking is try to checkthe cpu pins on the moathboard... will reply back
     
  14. akbaar

    akbaar Master Guru

    Messages:
    426
    Likes Received:
    55
    GPU:
    ASUS TUFF 3080 12Gb
    hmm tried to fix motherboard socket pins but no luck ...
    still 2x speed

    can you know where is the cpu pins location on the 6700k that handles pci-e ?
     
  15. mbk1969

    mbk1969 Ancient Guru

    Messages:
    15,541
    Likes Received:
    13,561
    GPU:
    GF RTX 4070
    IMO, to buy new motherboard is much more easier task.
     
    akbaar likes this.

  16. akbaar

    akbaar Master Guru

    Messages:
    426
    Likes Received:
    55
    GPU:
    ASUS TUFF 3080 12Gb
    yes true , its my last option master mbk1969

    my next step is to try the gpu on another pc, just to make sure its not the gpu! i hope its the mobo not the gpu haha
     
  17. norton

    norton Master Guru

    Messages:
    214
    Likes Received:
    56
    GPU:
    GTX970 G1 Gaming
    i had the same problem but i fixed it by cleaning the the GPU and PCI E slot the problem was the dust
     
  18. Calmmo

    Calmmo Guest

    Messages:
    2,424
    Likes Received:
    225
    GPU:
    RTX 4090 Zotac AMP
    Tricky issue, i have an issue my self with x8 on my zen2. Could be anything, CPU, mobo, gpu, or bad placement (user error).

    Clean the connectors, and with the mobo out of the case place it down again and check the the bus speed. (to remove the possibility that its the case/mobo mounting position causing the GPU to not be perfect sat on the slot)
     
  19. bobblunderton

    bobblunderton Master Guru

    Messages:
    420
    Likes Received:
    199
    GPU:
    EVGA 2070 Super 8gb
    First check that your power saving modes are off. Shut them all off, in BIOS, and in Windows; do the Windows ones first though. Save your old BIOS profile before you start messing around with things so that, if needed, you can come back to things later more easily and make it work again.
    You could have a bug in the PCI-express link state with power-saving feature. It's a feature in BIOS that cuts the amount of lanes during power save mode, as using the desktop obviously doesn't require much. You can disable this in BIOS usually.
    Likely somewhere in the PCI-E x4 link area is the issue with it, IF there's an issue with hardware and it's not software.

    Try GPU in another system. If good, then GPU is fine; if not, you have found your issue.
    Use the other slot* for GPU's (which is usually x8 electrical, x16 physical) if you have a second x16 slot or an open-ended x8 slot on your motherboard, to work-around this issue. If this does NOT work around the issue, it's either the GPU (see above test) or something in the cpu pins/socket or a defect with the cpu itself (unlikely but still remotely possible). It's almost always the motherboard, heck I've even seen a dead bug short things out for a while until it eventually dried up.
    You can pick an old mining motherboard up off ebay for 40$ USD at times, for Skylake chips, even if it's H110 it should do you just fine provided you can get a 10~12$ x4 link adapter card for your NVME drive if you need it. It's very possible one of the electrical PCI-E switches or capacitors went bad on the motherboard.
    *Using an x8 connection for your GPU vs x16 will not starve FPS beyond 2~3fps max, and only really be noticeable if you run out of VRAM and it has to page to system memory, which should almost never happen. So using the 2nd slot should be fine.

    If you hadn't taken a look at the CPU socket without daylight or the best interior light possible *AND* a magnifying glass, then you haven't really looked at it.
    If it is not the socket, could be just a fluke that the CPU went out in some way that broke one of the PCI-E lanes inside of it... that almost NEVER happens, there's about a 0.2% chance (one-fifth of one percent) of that ever happening that the CPU dies in some way unless it's massively overvolted AND overheated.

    PCI-E handshake goes like this: x1, x2, x4, x8, x16, in that order. It uses the highest level it can handshake on successfully and runs with it, aside of dropping link-state width when power-save modes are on. The processor and board need progressively stronger power as the link width increases.
    A poor quality or mis-shapen case can cause an issue with this handshake, as the contacts might not line up properly.

    I would not reckon that it's a power supply due to you not having other wacky or otherwise strange errors with the PC.
    Lastly, if you have some type of short in the secondary x16 slot, it's remotely possible it's throwing off something in the primary x16 slot. Normally this creates a black screen boot issue, or less destructively causes an x8 link vs x16 link on the main slot due to lane sharing, but I am only listing this out of an abundance of caution. If you've used liquid metal or similar CONDUCTIVE stuff, it could have oozed out from under the heatsink/fan/waterblock on your motherboard CPU socket and gotten into the slot near the x4 link connections on your primary gpu x16 slot, causing the handshake at x4 and above to fail, but admittedly this is a real rarity.

    Dust can most certainly cause this, too, so make sure you've tried the simple stuff before you throw in the towel, so to speak.

    If you don't feel like doing all this and you know it's not the video card (by testing video card in another computer), replace the motherboard with a used or refurbished model.
     
  20. Hog54

    Hog54 Maha Guru

    Messages:
    1,246
    Likes Received:
    68
    GPU:
    Asus Tuf RTX 3070
    Does it say the same thing in the bios?
     

Share This Page