Threadripper X399 system keeps randomly locking up

Discussion in 'General Hardware' started by Zepher, Sep 27, 2019.

  1. Zepher

    Zepher [H]ipster Replacement

    Messages:
    16,958
    Joined:
    Sep 29, 2001
    My friends son's X399 system has become unstable a month or so ago and they brought it to me to see if I could get it working.
    Specs: AMD 1950 threadripper, Gigabyte X399 Gaming motherboard, 4x8GB Gskill ram, EVGA 1080ti FTW3, EVGA 850 B3

    When he dropped it off it wouldn't power on, reseated all the connections, still wouldn't power on.
    Hooked up a new psu and it fired right up, but it locked up on the logon screen after 5-10 minutes.
    Grabbed a spare SSD and did a fresh Windows 10 install and the machine was stable, ran Furmark for 15 or so minutes with no issues, ended up just letting it play some movies for a few hours before I hit the sack.

    Next day I get the password and hook his SSD back up and I log on and it locks up after a few minutes. Rebooted and it locked up again after a few minutes.

    Did a clean install of Windows on his SSD and it locked up as soon as setup was finished and when the machine restarted it was asking for boot device. Went into the BIOS and the SSD was selected as the first boot device but it wouldn't boot anymore.
    Did another clean Windows install and it locked up while it was doing the updates. When it rebooted it was asking for a boot device again.
    Now I am thinking the SSD is flaky so I grab my SSD and hook it up and machine fires up and works great.

    I tell him what was going on with the SSD so he decided to get a 1TB NVME drive.
    We install the NVME drive and install Windows and some games on it and it seems to be working so he takes it home.
    Next day he lets me know that it is locking up again either on the desktop while idling or when playing games.

    I had him remove all but 1 ram stick and move it between each ram slot if the machine crashes. crashed with each of the 4 sticks in all of the slots.

    He brought it by last night and tested some games to show me what was going on, Overwatch worked with no issue, 3D Mark ran fine, Furmark ran fine. GTA5 Online and this Samurai game locks up the machine.
    Updated the BIOS, that didn't change anything.

    A few hours ago I put his GPU in my PC and it worked fine. Put my GPU in his PC and it crashed to desktop with GTA online and the Samurai game locks up the PC.
    Also, there are no blue or black screens, the PC just freezes showing whatever was on the screen.
    I also noticed that the GPU is really hot to the touch.
    I put his GPU back in ran the Samurai game since that game loads up fast, played it till the machine locked up and touched the GPU and it was just as hot as mine was in his machine.

    I set the GPU fans to 100%, fired up the game and it crashed a few minutes later and the GPU was still very hot to the touch.
    I also tried different PCIe slots on his board as well.

    Is it possible that a flaky motherboard could cause the GPU to run hotter than normal?

    What do you all recommend?
     
  2. primetime

    primetime [H]ardness Supreme

    Messages:
    6,048
    Joined:
    Aug 17, 2005
    Have you tried the gpu in a different slot? Could be something about that slot the doesn't agree with that card
     
  3. Zepher

    Zepher [H]ipster Replacement

    Messages:
    16,958
    Joined:
    Sep 29, 2001
    ya, tried different slots as well.
     
  4. primetime

    primetime [H]ardness Supreme

    Messages:
    6,048
    Joined:
    Aug 17, 2005
    Im a bit confused.....you have tried different gpu cards in the threadripper board? with exact same result?
     
  5. Zepher

    Zepher [H]ipster Replacement

    Messages:
    16,958
    Joined:
    Sep 29, 2001
    Ya, put my 1080ti FTW3 into his machine and it crashes as well..
     
  6. primetime

    primetime [H]ardness Supreme

    Messages:
    6,048
    Joined:
    Aug 17, 2005
    kinda looking like the mb has some kind of stability issue but it would sure be nice if you could conform it with a spare cpu. Have you took a good look at the pins and what not and made sure everything is good there?
     
  7. Zepher

    Zepher [H]ipster Replacement

    Messages:
    16,958
    Joined:
    Sep 29, 2001
    I guess I could look at the pins. I'll ask him if he had taken the processor out since building it.

    It ran for a few hours last night, played a couple rounds of Overwatch every now and then but then it did eventually lock up after playing around 4 back to back matches.
    GTA5 and that Samurai game still locks up the machine in minutes.
     
  8. primetime

    primetime [H]ardness Supreme

    Messages:
    6,048
    Joined:
    Aug 17, 2005
    i might try under clocking the GPU really just to confirm it has no effect. Could be a rare driver bug that shows up on that config?
     
  9. IdiotInCharge

    IdiotInCharge [H]ardForum Junkie

    Messages:
    12,063
    Joined:
    Jun 13, 2003
    - bent CPU pins
    - motherboard is whack
    - PSU issues (which if tested, points at motherboard being whack)
    - board / CPU hate the memory
     
  10. Zepher

    Zepher [H]ipster Replacement

    Messages:
    16,958
    Joined:
    Sep 29, 2001
    This is his machine. it's ran fine since he built it when the Threadripper came out. No hardware changes at all.
    Machine was overclocked and he put it to stock when it became unstable a month ago and it hasn't been stable since.
    IMG_3507.JPG
     
  11. Zepher

    Zepher [H]ipster Replacement

    Messages:
    16,958
    Joined:
    Sep 29, 2001
    Running Furmark and the external temp on the GPU gets to this temp but it's still running and hasn't crashed yet after 10 minutes.

    IMG_3527.JPG

    EDIT:
    it's actually 81.5 right now
     
  12. primetime

    primetime [H]ardness Supreme

    Messages:
    6,048
    Joined:
    Aug 17, 2005
    that seems fine to me....furmark doesn't always find instability that certain games might...but its a start
     
  13. Zepher

    Zepher [H]ipster Replacement

    Messages:
    16,958
    Joined:
    Sep 29, 2001
    I guess I can rule out the GPU getting too hot and crashing the machine as it's been running over 30 minutes at 82C.
     
  14. pendragon1

    pendragon1 [H]ardForum Junkie

    Messages:
    14,332
    Joined:
    Oct 7, 2000
    try bumping the ram and cpu voltages.
     
    Brian_B likes this.
  15. primetime

    primetime [H]ardness Supreme

    Messages:
    6,048
    Joined:
    Aug 17, 2005
    this is exactly my thoughts.....might just be a temperamental setup that needs exact voltages tweaked
     
  16. Zepher

    Zepher [H]ipster Replacement

    Messages:
    16,958
    Joined:
    Sep 29, 2001
    not related to the issue at hand, but I found out that this card has audio out on the DVI port.
    Using a DVI to HDMI cable since this monitor only has HDMI and my PC that is normally hooked to it has a GTX285 with only DVI ports.
    I know back in the day only certain cards carried audio on the DVI port.

    pic was taken with my 1080ti in his machine
    IMG_3524.JPG
     
  17. Zepher

    Zepher [H]ipster Replacement

    Messages:
    16,958
    Joined:
    Sep 29, 2001
    Stock was 1.20v on the CPU and Ram, and I upped it to 1.25 on them and no boot.

    EDIT, 1.22 on ram and 1.23 on CPU causes windows to lock up on the logo when booting.
     
  18. primetime

    primetime [H]ardness Supreme

    Messages:
    6,048
    Joined:
    Aug 17, 2005
    i think thats on the low side right?

    even my cpu runs up to 1.5 volts (as needed) and ram at 1.35 just because
     
  19. Zepher

    Zepher [H]ipster Replacement

    Messages:
    16,958
    Joined:
    Sep 29, 2001
    Are these the right settings to be changing?

    IMG_3528.JPG
     
    primetime likes this.
  20. primetime

    primetime [H]ardness Supreme

    Messages:
    6,048
    Joined:
    Aug 17, 2005
    should be this....i would do some heavy searches for recommended bios voltage max/min levels and it wouldn't hurt one bit if they used the same mother board.
     
  21. pendragon1

    pendragon1 [H]ardForum Junkie

    Messages:
    14,332
    Joined:
    Oct 7, 2000
    yeah should be. maybe try upping one but not the other, ram first.
     
  22. primetime

    primetime [H]ardness Supreme

    Messages:
    6,048
    Joined:
    Aug 17, 2005

    Attached Files:

  23. Zepher

    Zepher [H]ipster Replacement

    Messages:
    16,958
    Joined:
    Sep 29, 2001
    I was playing with it some more today and it is stable with 1 or 2 ram sticks, tried all of them, one at a time. With 4 sticks it locks up and I noticed that 2 of the sticks were much much warmer than the other 2, these were on the left side of the CPU.

    IMG_3530.JPG
     
    primetime likes this.
  24. primetime

    primetime [H]ardness Supreme

    Messages:
    6,048
    Joined:
    Aug 17, 2005
    well that sorta imply s pendragon may have been right all along.....drop the speed to 2100 and relax the timings to like 18 or higher (a few notches up from docp) and increase voltage to 1.4 and i bet you cant make it crash

    regardless.....yea i think its to do with the ram. interesting how moving slots helps
     
    Last edited: Oct 1, 2019
  25. Zepher

    Zepher [H]ipster Replacement

    Messages:
    16,958
    Joined:
    Sep 29, 2001
    just for shits and giggles I ran Memtest and it passed everything, took over 6 hours though,
    IMG_3533.JPG IMG_3534.JPG

    Oh, and I talked with him again last night to find out exactly what he was doing before it got unstable.
    He built the machine and had it OC'd to 3.7Ghz which is a mild overclock, and then a about month ago he was doing some benchmarks and used the AMD tuner software to revert the OC to stock speeds for comparison.
    Ran benches at stock speeds and it was fine, fired up a game and it crashed, and kept crashing in all the games, and that is where we are at, lol.
     
  26. pendragon1

    pendragon1 [H]ardForum Junkie

    Messages:
    14,332
    Joined:
    Oct 7, 2000
    remove ryzen master, reset bios, start over.
     
  27. primetime

    primetime [H]ardness Supreme

    Messages:
    6,048
    Joined:
    Aug 17, 2005
    memory test proves absolutely nothing other than totally defective ram.....It will NEVER identify unstable ram thats randomly crashing....but hey keep up the good fight lol!:)
    Pretty much 95% of bsod any of my setups ever had would not show anything from memory test, but would in fact be fixed with the correct ram settings....
     
    Last edited: Oct 1, 2019
  28. Zepher

    Zepher [H]ipster Replacement

    Messages:
    16,958
    Joined:
    Sep 29, 2001
    Did a few fresh installs and updated and reset the BIOS. Haven't installed or used the Ryzen master program.

    I'm going to try the timing thing just have to figure where it is and change it, I never really messed with the ram settings since I am not really an overclocker, just use basic cpu overclocking.

    EDIT: I set it to 2133, CAS 18, and 1.4v and so far the game is running longer with the 4 sticks of ram in it.
    gonna let it play and see if it locks up.

    EDIT2: it locked up when I stopped to type the first edit.
     
    Last edited: Oct 1, 2019
    pendragon1 likes this.
  29. primetime

    primetime [H]ardness Supreme

    Messages:
    6,048
    Joined:
    Aug 17, 2005
    well you had to try it................i wonder if just upping the timings to like 20ish (like docp plus 4 for each number) maybe just try the speed at default DOCP.
    I just know whats always worked for me so far.......seems like i have seen a couple of crashing threads with people using 1080ti cards lately. I once had an issue with an amd driver that defaulted the dam gpu voltage to low and would crash every time unless i boosted the voltage with afterburner until they finally fixed the driver issue. I just do know memory test has never been much help unless its totally defective ram (it doesnt find instability cause it doesnt load/push the ram the way games and apps do)
     
  30. Zepher

    Zepher [H]ipster Replacement

    Messages:
    16,958
    Joined:
    Sep 29, 2001
    I was also going to shift the ram over to the next channel but the cpu cooler has the wires coming out the side and blocks that dimm slot right next to it.
     
    primetime likes this.
  31. name-cannot-change

    name-cannot-change n00b

    Messages:
    3
    Joined:
    Sep 29, 2019
    Any updates? I've been following this thread because I'm having similar issues
     
  32. Zepher

    Zepher [H]ipster Replacement

    Messages:
    16,958
    Joined:
    Sep 29, 2001
    he bought a new CPU that we are going to drop in on Friday, if it still has issues then we are going to RMA the motherboard I guess.

    what is your setup and is it also hard locking while gaming? This one doesn't reboot, it just freezes showing an image of the game where it froze till you power down the machine.
     
  33. name-cannot-change

    name-cannot-change n00b

    Messages:
    3
    Joined:
    Sep 29, 2019
    Small snippet from dxdiag:


    Code:
             Operating System: Windows 10 Pro 64-bit (10.0, Build 18362) (18362.19h1_release.190318-1202)
                     Language: English (Regional Setting: English)
          System Manufacturer: Micro-Star International Co., Ltd.
                 System Model: MS-7B92
                         BIOS: BIOS Date: 07/24/18 15:40:49 Ver: V1.00 (type: BIOS)
                    Processor: AMD Ryzen Threadripper 1950X 16-Core Processor  (32 CPUs), ~3.4GHz
                       Memory: 32768MB RAM
          Available OS Memory: 32660MB RAM
                    Page File: 16288MB used, 21491MB available
                  Windows Dir: C:\WINDOWS
              DirectX Version: DirectX 12
    And I have 2 NVIDIA GeForce GTX 1070 and 2 M.2 drives
     
  34. primetime

    primetime [H]ardness Supreme

    Messages:
    6,048
    Joined:
    Aug 17, 2005
    and what exactly is yours doing? and when? when did it start? what steps have you tried so far?
     
  35. Zepher

    Zepher [H]ipster Replacement

    Messages:
    16,958
    Joined:
    Sep 29, 2001
    Installed the new Threadripper and had no issue playing games with all 32 gigs in there.
     
    IdiotInCharge and pendragon1 like this.
  36. Zepher

    Zepher [H]ipster Replacement

    Messages:
    16,958
    Joined:
    Sep 29, 2001
    Update, ended up having to swap out the motherboard.
    When he got home he noticed that the machine only showed 24GB of Ram. He brought it back along with a spare set of ram and it looks like the right side memory slots stopped working right.
    Only 1 stick would show up on that side. I tried all different configurations with various results.
    Ended up sticking all 4 sticks on the left side till the new board arrived.

    Oh, and when installing the new board, I noticed there was an extra stand-off at the top center under the VRM's. That might have been the cause of the issue we were having.
     
    Last edited: Oct 12, 2019
    pendragon1 likes this.
  37. name-cannot-change

    name-cannot-change n00b

    Messages:
    3
    Joined:
    Sep 29, 2019
    When you say that might have been the cause of the issue you were having, do you mean to the RAM not showing up or the computer freezing?
     
  38. Zepher

    Zepher [H]ipster Replacement

    Messages:
    16,958
    Joined:
    Sep 29, 2001
    Probably caused both issues. it might have been shorting something on the bottom of the board, and with all the pressure from constant re-inserting ram sticks, it could have made the shorting worse, if it was shorting.