vega 64 crashes unless in power save preset

Discussion in 'AMD Flavor' started by d8lock, Aug 4, 2018.

  1. d8lock

    d8lock Limp Gawd

    Messages:
    410
    Joined:
    Feb 12, 2012
    I'm on my second asus strix vega 64. It shows the same problems as the other. On certain games (Battlefield 1, Mass Effect Andromeda, Madden 19) it will crash within 10 minutes of playing. The only "solution" to this seems to be send the card back again. Is there something I'm missing with this card? When it runs, it works great.

    ryzen2700x
    gigabyte x470 wireless 7
    gskill3200 32gb
    evga g2 1000w psu
     
    Last edited: Aug 7, 2018
  2. thebufenator

    thebufenator Gawd

    Messages:
    921
    Joined:
    Dec 8, 2004
    More details.

    New card?

    Stock settings, just changing the presets?

    Temps?
     
  3. d8lock

    d8lock Limp Gawd

    Messages:
    410
    Joined:
    Feb 12, 2012
    It seems to be the power limit. Right now I've gotten it to -15 and it's still stable. New card (second new card) stock settings, looks like highest temp was 71c. I've updated the bios and that doesn't seem to have helped. Both bios' on the card appear to be the same. As was the other card I sent back.
     
  4. d8lock

    d8lock Limp Gawd

    Messages:
    410
    Joined:
    Feb 12, 2012
    Memory temps look like they stay right around 75c. I have it at -5 and still stable. Core clock is boosting to 1440s (frequency in wattmanis still on dynamic stock). Memory, also stock, stays pegged at 945 in game. This is also the second system this card has been in . Other intel 3770k this one ryzen 2700x.

    edit: looks like -2 is the highest i can put the power limit.
     
    Last edited: Aug 4, 2018
  5. thebufenator

    thebufenator Gawd

    Messages:
    921
    Joined:
    Dec 8, 2004
    Sure its not the PSU?
     
  6. d8lock

    d8lock Limp Gawd

    Messages:
    410
    Joined:
    Feb 12, 2012
    I don't think it is. I had two 580s before this, then a fury before that. I did notice the VR SOC temp was insanely high. 91c when in heavy load.
     
  7. DrLobotomy

    DrLobotomy [H]ardness Supreme

    Messages:
    5,786
    Joined:
    May 19, 2016
    Did they actually send you back a different card? You DID write down the S/N and take a pic of the magic sticker, eh?

    You have the power cable/s plugged into card? Are you using the correct ones?
     
  8. d8lock

    d8lock Limp Gawd

    Messages:
    410
    Joined:
    Feb 12, 2012
    Two different cards. From different stores. And yes, the psu cables are the separate ones, not daisy chained.
     
  9. DrLobotomy

    DrLobotomy [H]ardness Supreme

    Messages:
    5,786
    Joined:
    May 19, 2016
    The chances of that are pretty damn slim.

    Try a diff PSU or a different slot on the motherboard if you have it.
     
  10. d8lock

    d8lock Limp Gawd

    Messages:
    410
    Joined:
    Feb 12, 2012
    The power supply is the only component which I haven't swapped out... sigh. I really hope that's not it. Will try and report back.
     
  11. trick0502

    trick0502 [H]ardness Supreme

    Messages:
    5,071
    Joined:
    Apr 17, 2006
    You need to make sure the pcie power cables go to different rails.
     
  12. d8lock

    d8lock Limp Gawd

    Messages:
    410
    Joined:
    Feb 12, 2012
    As of now, it appears it was a bad power cable to the PSU. So far stable on 0 power limit. I was this close to putting an old tx850 with a loud fan in. Whew. Maybe now I can finally oc this beast.

    I assume it is. Using separate ports off the PSU.
     
    Ocellaris likes this.
  13. d8lock

    d8lock Limp Gawd

    Messages:
    410
    Joined:
    Feb 12, 2012
    It was the power cable. That's crazy. I think I've only run across a bad power cable one other time. All my problems were from that! Such a huge relief.
     
  14. DrLobotomy

    DrLobotomy [H]ardness Supreme

    Messages:
    5,786
    Joined:
    May 19, 2016
    Call the place you took back the card and tellem its prolly ok.
     
  15. d8lock

    d8lock Limp Gawd

    Messages:
    410
    Joined:
    Feb 12, 2012
    The first place (amazon) sent it without a shipping box. It came in the retail Asus box. So I was going to send that back even if it did work.
     
  16. dexvx

    dexvx Gawd

    Messages:
    843
    Joined:
    Aug 14, 2002
    Daisy chained PCIe cables shouldn't matter for a stock card.

    I have a blower Vega 64 Limited, and can undervolt/downclock to 1.075V, 1552 Core, +50% power limit.
     
  17. d8lock

    d8lock Limp Gawd

    Messages:
    410
    Joined:
    Feb 12, 2012
    I've read a lot of people having problems by doing that. I didn't ever hook mine up that way so wouldn't know.
     
  18. cageymaru

    cageymaru [H]ard|News

    Messages:
    18,253
    Joined:
    Apr 10, 2003
  19. d8lock

    d8lock Limp Gawd

    Messages:
    410
    Joined:
    Feb 12, 2012
    Problem seems to be back. Though not as frequent. Went ahead and ordered a new psu.
    What's weird is that it will crash a game. Then I can immediately restart said game, and it will run fine for hours. I've mainly been testing this with Andromeda. Vampyr crashes after 10 minutes every time.
     
    Last edited: Aug 7, 2018
  20. primetime

    primetime [H]ardness Supreme

    Messages:
    5,184
    Joined:
    Aug 17, 2005
    AMD at times releases certain drivers that have issues with some cards and you get the issues very close to what your getting......try a much older driver like a 17.7 version just to rule out some kind of software bug thats effecting just those cards. Seems like it was about a year ago i went thru several driver releases that just didnt agree with my card or anyone else with the exact same cards

    i think some of the drivers get the gpu voltage all messed up on certain models
     
    d8lock likes this.
  21. KazeoHin

    KazeoHin [H]ardness Supreme

    Messages:
    7,326
    Joined:
    Sep 7, 2011
    A friend of mine had a similar issue. Turns out some PSU's have overly sensitive overcurrent protection and some Vega operations trip them.
     
    d8lock likes this.
  22. DigitalGriffin

    DigitalGriffin [H]ardness Supreme

    Messages:
    4,513
    Joined:
    Oct 14, 2004
    Is it me or is 71C on the memory chips seem high? Could it be the chip packaging issue seen with different manufacturers of HBM2? The heatsink wasn't sitting flush against some mfg's versions because the chip package was thinner.
     
  23. d8lock

    d8lock Limp Gawd

    Messages:
    410
    Joined:
    Feb 12, 2012
    If that's true, then avoid the psu I've got. It can run stress tests for hours yet crashes on games. The newest drivers have made the system completely crash now, instead of the device removed error. I sure hope the power supply arriving tomorrow ends all this madness.
     
  24. d8lock

    d8lock Limp Gawd

    Messages:
    410
    Joined:
    Feb 12, 2012
    So far the Seasonic psu seems to be holding up. Haven't done any major testing yet, but it hasn't crashed. I wonder if it would be a good idea to start a list of problematic psu's with vega 64, or just this particular asus model.
     
  25. DigitalGriffin

    DigitalGriffin [H]ardness Supreme

    Messages:
    4,513
    Joined:
    Oct 14, 2004
    As power supplies get older they get both nosier and their total power output drops. Sometimes by as much as 20% or more.

    If you ever view Kyles articles on testing older PSU's after the 10 year mark you will see how drastically they change. Even a few years can make a marked difference if you consistently run it near the edge (heat death) or not at all (Cap depolarization)
     
  26. d8lock

    d8lock Limp Gawd

    Messages:
    410
    Joined:
    Feb 12, 2012
    Not disagreeing with you, but the evga psu is less than two years old. G2 1000w. Not an entry level psu either. It also runs every other card perfectly fine. Only this vega causes issues.
     
  27. d8lock

    d8lock Limp Gawd

    Messages:
    410
    Joined:
    Feb 12, 2012
    Crashed again. I'm now pretty sure this is all caused from driver issues. Going back to 18.5.1 to see if it's stable there.
     
  28. DigitalGriffin

    DigitalGriffin [H]ardness Supreme

    Messages:
    4,513
    Joined:
    Oct 14, 2004
    I would compare your mem chip temps to what the review sites are saying.

    Are you running standard bus speed?
     
  29. d8lock

    d8lock Limp Gawd

    Messages:
    410
    Joined:
    Feb 12, 2012
    Yeah I didn't change any bus settings. I did however have it running on the turbo preset in wattman. After putting 18.5.1 on I can just end the application that crashes instead of an unresponsive system with black screen. Going to keep it on balanced until I figure this out. On the bright side, I have a backup power supply now.
     
  30. DigitalGriffin

    DigitalGriffin [H]ardness Supreme

    Messages:
    4,513
    Joined:
    Oct 14, 2004
    Do you have more than 1 PCIe 16 slot?
     
  31. d8lock

    d8lock Limp Gawd

    Messages:
    410
    Joined:
    Feb 12, 2012
    Yeah, this is the motherboard I'm using. And yes, I have tried running it in a different slot. I'm thinking it's software related.
     
  32. primetime

    primetime [H]ardness Supreme

    Messages:
    5,184
    Joined:
    Aug 17, 2005
    I can guarantee it's software glitch just as mine had for a few months....they sorted it out eventually but I had to stay on older drivers that didn't have the glitch with my cards firmware....if we only had 1 version of Vegas this couldn't happen
     
  33. d8lock

    d8lock Limp Gawd

    Messages:
    410
    Joined:
    Feb 12, 2012
    Which drivers are you on? 18.5.1 I think is the furthest back I've gone.
     
  34. primetime

    primetime [H]ardness Supreme

    Messages:
    5,184
    Joined:
    Aug 17, 2005
    im running latest since they work fine FOR ME.....but you gotta do the leg work yourself...do some searches online and at least try a known good driver for your card.......learn how to use afterburner osd's and check out the gpu voltage on one driver versus another......when i was having issues i finely bothered to see what was going on and low and behold the dam driver was super under volting my card....go figure it was crashing
     
  35. d8lock

    d8lock Limp Gawd

    Messages:
    410
    Joined:
    Feb 12, 2012
    I haven't been using afterburner on this card since it had all wrong readings on my 580s. Maybe I'll go back to it. At least I can now put the computer back together. 18.5.1 so far is the best, but it's also the oldest I've tried. Glad to hear I'm not the only one that's dealt with this.
     
  36. ccityinstaller

    ccityinstaller 2[H]4U

    Messages:
    3,426
    Joined:
    Feb 23, 2007
    I'd be willing to bet $100 the "driver" itself was not undervolting your GPU. You most likely had old/corrupted PowerPlay tables in the registry that were playing havoc with your particular card.

    As far as the "if we only had 1 version of VEGA we would not be having this issue" I am sure what you are referring too unless you mean the fact that different vendors were used to pack the GPU die and HBM stack?
     
  37. primetime

    primetime [H]ardness Supreme

    Messages:
    5,184
    Joined:
    Aug 17, 2005
    as in different cards like xfx and or asus uses different firmware possibility....unless they got away from doing such things with vegas....the rx580s/480s certainly did....and no it wasnt power play on my pc alone.....pretty much everyone with xfx 580s (or the same models) were Crashing with the exact same drivers...like i said it was easy to see with afterburner what was happening (when i bothered to check)
     
  38. d8lock

    d8lock Limp Gawd

    Messages:
    410
    Joined:
    Feb 12, 2012
    Crashes on 17.10 drivers too. The best I can make it is so that it crashes the game within 5 min, then after opening it back up it works fine. Not an acceptable solution for me, so I will be sending this back and using a 1080 ti.
     
    Last edited: Aug 10, 2018 at 1:43 AM
  39. d8lock

    d8lock Limp Gawd

    Messages:
    410
    Joined:
    Feb 12, 2012
    Thanks to newegg's replacement only policy for this item, I guess I won't be switching. Sure hope I get a card that's not a turd. Third time's the charm.
     
  40. primetime

    primetime [H]ardness Supreme

    Messages:
    5,184
    Joined:
    Aug 17, 2005
    id say you should be able to get a different card that works....cause that rules out faulty drivers as well.....i would try a FRESH/NEW install of windows just to be 100% sure its a hardware issue....but otherwise not much else you can do
    Well you could sell the RMA as unopened/new....it's an option at least....if you open it and it's still faulty...well then you should tell any buyer what's up....I'm guessing you already was checking for newer firmware?