Crashes during gaming with solid color on screen

Discussion in 'AMD Flavor' started by blackmomba, Sep 25, 2019.

  1. blackmomba

    blackmomba [H]Lite

    Messages:
    119
    Joined:
    Dec 5, 2018
    got my card back from RMA and I still have the crashing during gaming with a solid color on the monitor. A lot of googling led me to believe it was a problem with the card

    2700x
    Prime X470-PRO (latest BIOS)
    Ripjaws V DDR4 3200 C16 16-18-18-36-58
    Strix Vega 64 (fresh from RMA)
    Corsair RM750 (was EVGA 850 B3)

    Tried loosening timings, thinking about turning off PBO to see if that helps.

    I've got regular peripherals and a wireless controller receiver with 5 HDD and 1 nvme drive. Could I be teetering on the edge of what I would need for power ?

    Any ideas?
     
  2. THUMPer

    THUMPer 2[H]4U

    Messages:
    3,112
    Joined:
    May 6, 2008
    What driver version? Is the vega overclocked?
     
  3. blackmomba

    blackmomba [H]Lite

    Messages:
    119
    Joined:
    Dec 5, 2018
    19.9.2

    Card is not at all oced. Bone stock. Usually sits between 1450 -1520 and afterburner shows around 280w consumption during full load
     
  4. THUMPer

    THUMPer 2[H]4U

    Messages:
    3,112
    Joined:
    May 6, 2008
    How many monitors? Which ones? HDMI, DP?
     
  5. blackmomba

    blackmomba [H]Lite

    Messages:
    119
    Joined:
    Dec 5, 2018
    3 monitors, 1x DP at 4k and 2x HDMI at 1080p. All samsung but different models

    p9QxbKK.png
     
  6. THUMPer

    THUMPer 2[H]4U

    Messages:
    3,112
    Joined:
    May 6, 2008
    whew. ok. what if you just go down to 1 monitor?
    Does the whole pc crash or just the driver? What if you unplug them and plug them back in once it goes green. Does the picture come back?
     
    SticKx911 likes this.
  7. mvmiller12

    mvmiller12 Gawd

    Messages:
    919
    Joined:
    Aug 7, 2011
  8. blackmomba

    blackmomba [H]Lite

    Messages:
    119
    Joined:
    Dec 5, 2018
    Picture never comes back. Either it reboots or the video never comes back. I haven't tried unplugging so I'll try that next time. Whats annoying is that I've found no way to reproduce the issue consistently.

    I've had it happen during Hitman 2016, PoE, Heroes of the Storm, Mafia 2....
     
  9. mvmiller12

    mvmiller12 Gawd

    Messages:
    919
    Joined:
    Aug 7, 2011
    So, nothing super-demanding then. zDo you have a spare power supply? If you do, try connecting just the PCI-E Auxilliary power plugs from that to the video card, and jump the power supply so that it runs. This will pull pretty much all of the strain off your regular power supply for running the card. If the problem goes away, there lies your answer :).

    The truth is this: You've already RMA'd the card and tried different slots. Unless they sent you back the exact same card you sent in, the card should be presumed OK. The mainboard * could * be the culprit, but generally that is not the kind of failure you see with those, and testing an alternate slot further reduces that likelihood. That realistically leaves the monitor itself, the connecting video cable, and the power supply as possible culprits. Video cables rarely fail, and if the issue is the monitor itself, I would expect the issue to be more consistent than "only when playing games." This leaves the power supply as the more likely suspect.

    Edit, I HAVE had the green screen issue with my Vega 64, but it only ever happens when installing/updating drivers, and a simple reboot squares it.
     
    blackmomba likes this.
  10. blackmomba

    blackmomba [H]Lite

    Messages:
    119
    Joined:
    Dec 5, 2018
    Thanks bub

    Didn't think of using a dedicated ps. If only I didn't rush to sell the 850 B3 i had hehe

    I honestly don't think I have any others that would work, all I've got are some old P4 PSUs

    I will also try to consolidate drives because 5 seems like a lot for nothing. Should help reduce the draw

    Thanks again
     
  11. DooKey

    DooKey [H]ardForum Junkie

    Messages:
    8,250
    Joined:
    Apr 25, 2001
    I had that problem with my RX570 and my 3900X system. I also use the same RAM that you have and my problem went away when I ran the RAM at 2933.
     
    blackmomba likes this.
  12. blackmomba

    blackmomba [H]Lite

    Messages:
    119
    Joined:
    Dec 5, 2018
    Thanks for the feedback. I'm still pursuing the "not enough power" route for the moment because I used a Rx480 Nitro while waiting for vega to be RMAed and never had a single issue.

    I got myself a killawatt and saw that at full load (prime small fft and furmark 4k stress test) everything together pulls 750w almost exactly.

    So IMO the RM750 is borderline enough while the B3 should've been enough.. not sure why it wasn't. I ended up selling it and the guy hasn't complained

    I remembered that I had a 850 G3 in a second machine. I put it in last night and let furmark run all night. So far so good. played 2 games of hots this morning and no crash

    I'm gonna keep furmark and prime running all day to see if I can provoke the problem still
     
    AlphaQup, c3k, AceGoober and 2 others like this.
  13. mvmiller12

    mvmiller12 Gawd

    Messages:
    919
    Joined:
    Aug 7, 2011
    Are you testing on the Vega or on the RX480?
     
  14. blackmomba

    blackmomba [H]Lite

    Messages:
    119
    Joined:
    Dec 5, 2018
    Vega for sure

    GPU temp around 67C (76% Fan speed)
    GPU mem temp 74C

    CPU Tctl 94
    CPU Tdie 84

    CPU seems pretty high....
     
  15. mvmiller12

    mvmiller12 Gawd

    Messages:
    919
    Joined:
    Aug 7, 2011
    Temp looks normal for the stock 2700x cooler running Prime95 on "FRY ME!!!!!!" mode :). It's not going to hurt the chip, but you might get some thermal throttling at that temperature. What cooling solution are you using for the CPU?
     
    AceGoober and blackmomba like this.
  16. auntjemima

    auntjemima [H]ardness Supreme

    Messages:
    5,558
    Joined:
    Mar 1, 2014
    5 drives require nearly no power. I wouldn't worry about consolidating for the possible 100w you might save (and that's a very generous number).
     
    blackmomba likes this.
  17. blackmomba

    blackmomba [H]Lite

    Messages:
    119
    Joined:
    Dec 5, 2018
    Nothing special This fucker https://noctua.at/en/nh-u12s-se-am4

    Thanks a lot
     
    auntjemima likes this.
  18. blackmomba

    blackmomba [H]Lite

    Messages:
    119
    Joined:
    Dec 5, 2018
    So after like 18 hours running the machine at full load as described above, still didn't yield any negative results.

    Within a half hour of playing Hots, the machine crashed (I know hots sucks)

    I am now testing in the second x16 slot. I dont want to have to RMA my board so part of me hopes the machine crashes again but obviously I wanna find the root cause

    Anyway, I should've tried the second slot straight off the bat but I'm older now and it's been years since I needed to troubleshoot so forgive me

    I only keep updating this thread because despite the problem being well documented, the solutions seem to vary quite a bit. So I'm documenting mine in hopes some poor soul is aided by my pain
     
    AceGoober likes this.
  19. mvmiller12

    mvmiller12 Gawd

    Messages:
    919
    Joined:
    Aug 7, 2011
    The bigger question is this:

    Did you reboot the computer in between completing your testing and trying to game? Hell, MY computer is wonky as shit after an extended Prime95 run and I have to reboot it to restore stability although all of my stress testing has passed. I also have to reboot after completing large batches of x264 encodes or the computer behaves... strangely. This has been true for every PC I have ever owned. I believe this problem is located more firmly in the Windows wheelhouse than in the hardware one, though. I've not heard of LInux users having this kind of problem, but it is very common in Windows.
     
  20. blackmomba

    blackmomba [H]Lite

    Messages:
    119
    Joined:
    Dec 5, 2018
    I see what you're saying. I however prefer that my machine remain stable at all times, more so when running stock settings. I also find that W10 and even 8.1 don't need to be restarted as often as their predecessors did

    I used to run folding@home on a 1055t in 2009-2010. At the time I was doing -bigadv work units which were normally only assigned to 8 core chips. Under linux we used to fake an extra two cores because a heavily overclocked 1055t could actually complete the work unit right before the deadline and thus be awarded points. If my machine became unstable while crunching, the downtime would cause me to miss the deadline, lose the points and time spent. My team would've kicked me out. So I spent a lot of time stress testing, typically during a 3 day period (roughly the time it took to complete a -bigadv wu). I still use this chip today in a machine that runs part of a kubernetes cluster. Stability in my workstations is incredibly important to me.

    While using the second pcie slot, I played a whole bunch of games over the last 10 days or so; kcd, warframe, division 1, poe and hots. Machine crashed 3 times, did not reboot on it's own. Crashed once at the desktop without anything really running except my browser and Im watching plex. That sucked

    I don't have many hardware utilities installed; afterburner, icue, got the steelseries driver and utility. I've got some other services running but nothing I would think that could conflict with the display. I've done the DDU dance a couple times already but I'm willing to try again. I'm hesitant to wipe and reinstall the bare essentials but I guess it's my next step. In the meantime I'm reaching out to asus for feedback and to see if I can get yet another RMA.
     
  21. pendragon1

    pendragon1 [H]ardForum Junkie

    Messages:
    15,526
    Joined:
    Oct 7, 2000
    try a different psu yet? try running with all the background stuff disabled.
     
  22. blackmomba

    blackmomba [H]Lite

    Messages:
    119
    Joined:
    Dec 5, 2018
    I've tried three so far

    EVGA 850 B3
    Corsair RMx750
    currently EVGA 850 G3
     
    pendragon1 likes this.
  23. Furious_Styles

    Furious_Styles [H]ard|Gawd

    Messages:
    1,543
    Joined:
    Jan 16, 2013
    After reading the thread it does seem like a mobo problem. You have pretty much ruled out the PSU, CPU, and GPU (rare to get a bad one from RMA) only thing left really is board or mem. Have you tried removing everything but the barest of necessities just to see?
     
  24. filip

    filip [H]ard|Gawd

    Messages:
    2,031
    Joined:
    Aug 15, 2012
    My personal experience, mem problems can cause bsod and crashes in game (you can have a good and bad stick or just shit memory from the factory, when it is loaded it will fail), a gpu going bad can cause problems, mb was only ever a total failure.

    Check that memory sounds like the problem.
     
  25. Zepher

    Zepher [H]ipster Replacement

    Messages:
    17,155
    Joined:
    Sep 29, 2001
  26. crazycrave

    crazycrave Gawd

    Messages:
    628
    Joined:
    Mar 31, 2016
    It worked for me with RX 5700 acting like I had Witch in my drivers , I have CX 650 M and 8 pin connection is on it's on rail alone , 6 pin connection is on the other rail alone and now 100 % stable and no more anything ,
     
  27. ThreeDee

    ThreeDee [H]ardForum Junkie

    Messages:
    10,720
    Joined:
    Sep 5, 2001
    I rma'd a GTX 1070 to EVGA ... they sent back what appeared to be the same card with the same exact problem .. I rma'd that back to them ..again .. after talking to one of their techs and asking if they even test these things before they send them out .. the next card I got back worked .. so RMA screw ups happen .. just throwing that out there. :)

    ..maybe try the card in a different box and see what happens
     
  28. blackmomba

    blackmomba [H]Lite

    Messages:
    119
    Joined:
    Dec 5, 2018
    I haven't no. Haven't had time. I ran memtest for a few hours and it came back clean. My sticks are not on the QVL though. I clocked them down to the 2133MHz with auto timings and had the same issue. I used the Ryzen DRAM calculator for timings ever since I started with Ryzen on my 1700 (same RAM) and that build worked well.

    Thanks, not exactly my scenario but I'll keep it in mind

    Have had this since the start

    Thanks. I've tried to chat with Asus a couple of times. Wait times are wild.

    I don't really have another machine that can take this card at the moment. I have a mATX biuld that this card won't fit in and the other build is running headless with Ubuntu.

    The worst part, the absolute worst part is not being able to reproduce the problem. I can run stress tests and benchmarks all day long and the machine stays stable. Only after a couple of hours of fucking hots does it crash.. oh and randomly at the desktop.

    I wanted to see if it was my card boosting too high or something like that so I turned on AB logging. This is what things were like right before a crash:

    upload_2019-10-21_8-51-21.png

    I think I'm going to get a board just to throw something new into the mix.
     
  29. Furious_Styles

    Furious_Styles [H]ard|Gawd

    Messages:
    1,543
    Joined:
    Jan 16, 2013
    I see nothing wrong really in that chart. GL with the board that is certainly my best guess at this point. Asus also generally is terrible for support but hopefully you get lucky.

    These problems can definitely be a real PITA to isolate sometimes, especially if you don't have a bunch of spares lying around. I had one recently where my monitor USB hub was causing random crashes (sometimes days, sometimes weeks of uptime) and I was thinking it was my OC. I finally realized what it was when I turned windows sounds back on and kept hearing the disconnect/reconnect sound. No problems since completely disconnected the USB from the monitor.
     
  30. blackmomba

    blackmomba [H]Lite

    Messages:
    119
    Joined:
    Dec 5, 2018
    You know what... Speaking of USB.. I remembered that I've been having some trouble with my mouse where I need to unplug and plug it back it for it to work, I think it even happened once after one of these crashes. Using a SteelSeries Rival 710. Might be something pointing towards the board too

    Thanks!
     
  31. ThreeDee

    ThreeDee [H]ardForum Junkie

    Messages:
    10,720
    Joined:
    Sep 5, 2001
    random crashes can be a bear .. my daughters mini-ITX setup had been running fine for almost a year and then random lock ups/BSOD .. tracking down a bunch of errors in the Event Log finally pointed to the Intel WiFi card .. drivers from ASRock were crashing and went to Intel and got their updated drivers and all has been good .. sift through your Event Log and see if it says anything if you haven't already
     
  32. pendragon1

    pendragon1 [H]ardForum Junkie

    Messages:
    15,526
    Joined:
    Oct 7, 2000
    so i just updated my system(board, chip, ram) on the weekend and was having all sorts of random crashing. turns out i needed to RTFM to find which RAM slots should be used, put the sticks in the correct slots and it now seems rock solid. maybe check on that.
     
    blackmomba likes this.
  33. blackmomba

    blackmomba [H]Lite

    Messages:
    119
    Joined:
    Dec 5, 2018
    Swapped the Prime x470-Pro for a x470 Gaming plus and after a few days had the issue again. Gonna swap RAM next, then nvme drive. After that there will be nothing left to replace

    A positive that came of this is that this little MSI board boosts the 2700x much higher than the prime did. I see a max of 4.35GHz now whereas before it wouldn't break 4.2GHz. happy with that
     
  34. blackmomba

    blackmomba [H]Lite

    Messages:
    119
    Joined:
    Dec 5, 2018
    IMO RAM is fine so I'm not gonna bother. Can't pass this many rounds of mem test and still crash randomly so I'm not gonna waste my time there

    I already sold my prime board

    I went back to Asus and chatted with a rep yesterday who confirmed that my card wasn't replaced but merely repaired. They are offering to repair AGAIN because apparently they don't have any inventory ATM. I'm asking that they replace the card but they don't want to. So I escalated and I'm waiting for someone to reach out. I don't know how they can justify another "repair' when they tried one already and seemingly sent me back the same problem

    Any advice ?
     
  35. pendragon1

    pendragon1 [H]ardForum Junkie

    Messages:
    15,526
    Joined:
    Oct 7, 2000
    nope. i dont think we have an asus rep on the forum, i could be wrong and someone will correct if so. just stand your ground and demand a replacement.
     
    blackmomba likes this.
  36. ThreeDee

    ThreeDee [H]ardForum Junkie

    Messages:
    10,720
    Joined:
    Sep 5, 2001
    the already didn't fix it once .. keep asking for the next person up the food chain until they replace your card .. not "fix" it ..
     
  37. mvmiller12

    mvmiller12 Gawd

    Messages:
    919
    Joined:
    Aug 7, 2011
    My concern is that Asus is almost certainly asking you to pay for shipping it back to them. If nothing else, that should be a point of contention because you already had to do this the first time and it's THEIR screwup, not yours.
     
    blackmomba and pendragon1 like this.
  38. blackmomba

    blackmomba [H]Lite

    Messages:
    119
    Joined:
    Dec 5, 2018
    So I should be getting my card back from it's second repair today or tomorrow. They offered to pay for shipping so I went ahead. I've been on a Sapphire 7870 for a week without any problems (except for a desperate lack of VRAM.)

    I've dealt with Gigabyte and Sapphire RMA in the past and I'm sure they're all kind of similar, but I have to say overall I'm pretty unimpressed with ASUS customer service. They lack the basics. Crazy wait times, very little communication, if you ask a question that you couldn't answer yourself using self serve tools, it's always a 1-2 business day delay before you get an response. I was told someone would follow up, didn't happen. Zero transparency, so robotic. At my preferred price point, there are usually better products available anyway so I'll probably go with another AIB partner next time around
     
  39. blackmomba

    blackmomba [H]Lite

    Messages:
    119
    Joined:
    Dec 5, 2018
    Got a card back a couple weeks ago. Card was replaced. Different card, same issue.

    I got some value ram Aegis 2400mhz and after a few days had the issue at the desktop.

    So I'm down to
    a) CPU
    b) boot drive (intel 760p)

    So I picked up a 2200G. 75$ USD is a great price for a quad core with graphics. I'm going to install it later. If it turns out to be my 2700x, it'd be the first chip that I've ever had that became faulty on its own.
     
  40. THUMPer

    THUMPer 2[H]4U

    Messages:
    3,112
    Joined:
    May 6, 2008
    Are you running XMP profile on your RAM?