Crash, reboot cycle. GPU?

Discussion in 'Videocards - AMD Radeon' started by shimyns, Oct 8, 2015.

  1. shimyns

    shimyns Guest

    Messages:
    1,665
    Likes Received:
    1
    GPU:
    SAPPHIRE r9 290x Tri-X OC
    hi all,
    This started happening last night playing Nosgoth. Black screen or freeze, then reboot, cycles a few times and enters windows.

    Haven't changed any software or hardware. Temps look fine. Don't have any other games to try atm.

    Where do you think I should begin troubleshooting? Suggestions please and thank you. Also is there an easy free prog to test GPU stability? Does this sound GPU-related? RAM? CPU? Mobo? Really not in the mood for this treasure hunting...
     
  2. Undying

    Undying Ancient Guru

    Messages:
    25,343
    Likes Received:
    12,754
    GPU:
    XFX RX6800XT 16GB
    Remove overclock from everything, try stock clocks cpu/ram/gpu and see if problem persists.
     
  3. shimyns

    shimyns Guest

    Messages:
    1,665
    Likes Received:
    1
    GPU:
    SAPPHIRE r9 290x Tri-X OC
    Thanks,
    The only thing I have OC'ed is my CPU and I've had the same settings for 4 years with zero stability issues. It doesn't crash when I run Prime95 on blend test. So, I'm thinking it's more likely to be GPU related...

    EDIT: it did freeze and reboot somewhat randomly during Prime95 (no blue screen, just freeze and reboot).
     
    Last edited: Oct 11, 2015
  4. shimyns

    shimyns Guest

    Messages:
    1,665
    Likes Received:
    1
    GPU:
    SAPPHIRE r9 290x Tri-X OC
    I don't understand why when it crashes and reboots it enters a boot cycle where it reboots itself a 2-3 times before entering windows until it is finally able to...

    I was running Furmark just now and temps were absolutely fine and it suddenly blackscreened and reboot. Don't know if that's by chance or not.

    EDIT: I've increased some voltages and it might be stable now...
     
    Last edited: Oct 8, 2015

  5. PrMinisterGR

    PrMinisterGR Ancient Guru

    Messages:
    8,125
    Likes Received:
    969
    GPU:
    Inno3D RTX 3090
    Any chance to check another PSU?
     
  6. shimyns

    shimyns Guest

    Messages:
    1,665
    Likes Received:
    1
    GPU:
    SAPPHIRE r9 290x Tri-X OC
    That would have been my first guess but it's less than a year old and definitely doesn't have a lot of load on it so I don't think it's the PSU.

    I increased my vcore, qpi, and pll (i think) by one notch and it didn't reboot during furmark or playing, although I haven't tested it very extensively. I'm hoping it just needed some more juice as mobo and cpu are aging. When I have time I'll turn each of those three voltages back down and see which is the culprit.
     
  7. AsiJu

    AsiJu Ancient Guru

    Messages:
    8,810
    Likes Received:
    3,369
    GPU:
    KFA2 4070Ti EXG.v2
    The reboot cycle could be your mobo testing RAM compatibility settings so RAM could be the culprit.

    Try booting your system with just one module at a time.

    edit: ah, saw you got rid of crashing atm. Well, if your system starts doing that again, you could test the RAM then.
     
    Last edited: Oct 10, 2015
  8. shimyns

    shimyns Guest

    Messages:
    1,665
    Likes Received:
    1
    GPU:
    SAPPHIRE r9 290x Tri-X OC
    No, I thought I had but it's still happening. Unrelated to game play. I booted up my computer and the monitor was intermittently blacking out every few seconds, even while I was just in the BIOS menu (not crashing but blinking on and off)!

    At this point, I was sure it was the GPU or the HDMI cable. Otherwise, how/why could it be blinking like that in the BIOS menu?! I switched the cable, no help. I took off all of my CPU overclocks and set everything to default clocks and volts. even my RAM is at 1066 aven though they're rated for 1600. I booted up and everything seemed fine and my conclusion was that my CPU overclock must have been the reason; some component just became degraded and unstable.

    However, just now (a day later) I turned on my PC and it needed to automatically boot cycle 3 times before it made it into windows. So, now, I'm at a loss for what's causing this.

    Bottom line is that nothing is overclocked here (even though I had not made any software or hardware changes prior to this) and I have this boot issue!

    Any pros out there have suggestions?

    ps- Is it at all possible that my monitor is somehow causing some electrical short and that it's not even the PC that is the problem?
     
    Last edited: Oct 11, 2015
  9. shimyns

    shimyns Guest

    Messages:
    1,665
    Likes Received:
    1
    GPU:
    SAPPHIRE r9 290x Tri-X OC
    So, it looks like it reaches a message saying "detecting dram size" and then reboots. This happens 3 times. On the fourth time, it boots into windows. I'm researching this now...

    EDIT: I've seen a bunch of people get this problem and it obviously means memory issues. I'm tested each module by itself in each of the slots and it booted but it seems like when I set the speed to 1600 Mhz (like it's been for the last 4 years), I get this boot issue. I've reduced the multiplier to 6 and am now at 1200 Mhz and it booted without issue. I have yet to see if it remains stable.

    I don't know if it is the mobo or one of the memory modules that has degraded.. because I suppose that it could be either of these things that is the source to the problem (if this is in fact the problem, i.e., memory speed).
     
    Last edited: Oct 11, 2015
  10. AsiJu

    AsiJu Ancient Guru

    Messages:
    8,810
    Likes Received:
    3,369
    GPU:
    KFA2 4070Ti EXG.v2
    ^ so it seems trying to run your memory at its highest speed @ dual channel mode causes the issue.
    I had an almost identical problem with bent CPU pins on the mobo (most likely ones related to memory controller on the CPU). You could check the socket to make sure its ok.

    Though in your build the memory controller is on the motherboard I think. The modules are most likely fine if both worked independently. Could be your motherboard (or CPU) then causing this.
    I would think mobo is most likely, as running stock clocks didn't help and the AFAIK fact the memory controller is on the mobo.

    Try increasing DRAM voltage slightly if you can and check if you can boot up normally with normal memory config.
     
    Last edited: Oct 11, 2015

  11. shimyns

    shimyns Guest

    Messages:
    1,665
    Likes Received:
    1
    GPU:
    SAPPHIRE r9 290x Tri-X OC
    Thanks for your suggestions. These are actually triple channel modules (and 3x 2Gb sticks). I've heard about bent pins causing related issues but I seated my CPU 3 years ago and it hasn't been disturbed since then. Also, according to you, the memory controller is on the mobo (IDK) so I'm hoping it's not related to that.

    Currently, I have my same old CPU overclock (200 blck x 20 with dynamic vcore and my regular increased dll and qpi voltage). My memory is at 1.6v as it has been. The only thing I've done is reduce the spd multiplier to 6 in order for the ram to run at 1200 instead of 1600. It "seems" stable and I haven't gotten the infamous triple-reboot "detecting dram size" issue, but I really haven't stressed the pc since.

    I can try to increase the dram v to 1.62 or 1.64, I guess. However, I've read in a few places that QPI voltage can be the underlying reason for ram issues similar to these. So, I'm considering increasing my QPI voltage from its current 1.35v.

    Honestly, I don't feel like tinkering with this; I end up spending hours at a time troubleshooting and I don't really have the patience for it anymore (at this age, haha). Who knows? Maybe more QPI v is the solution...

    Thanks for your ideas and if you have anymore insight, keep it coming!
     
  12. AsiJu

    AsiJu Ancient Guru

    Messages:
    8,810
    Likes Received:
    3,369
    GPU:
    KFA2 4070Ti EXG.v2
    Give increasing the QPI voltage and DRAM voltage a try, increase just slightly and see if it makes a difference. If not, don't try "maxing" out either ;)

    BTW if you haven't touched your CPU/cooler in 3 years, it might be time to clean the cooler and reapply thermal paste...
    while at it you can check the socket and re-seat the CPU (with any luck this issue is just a bad contact somewhere). Just a suggestion.

    Report back after trying the voltages, if you like.
     
    Last edited: Oct 11, 2015
  13. shimyns

    shimyns Guest

    Messages:
    1,665
    Likes Received:
    1
    GPU:
    SAPPHIRE r9 290x Tri-X OC
    Yeah, I'm not going to try maxing them out, trust me. I won't go past 1.64 for ram. QPI is at 1.35. I think I'd be safe up to 1.55? I would say that if it IS these voltages, it shouldn't take more than 1 or 2 increments to get them back to working as they have been up until now.

    As for the CPU, I really don't feel like pulling it apart. I'll have to be in the "mood" to spend the time to do that (take it apart, clean the CPU thoroughly, apply new paste, and close it all back up). I mean, right now my temps are fine. Real temp has idle cpu at ~35 for all 4 cores and not above 70 when blend testing with Prime95 (even with HT enabled), so I think it's still seated well. And yeah, I let it blend test for an hour or so and it was fine.

    Will report back when I try the qpi voltage.
     
  14. AsiJu

    AsiJu Ancient Guru

    Messages:
    8,810
    Likes Received:
    3,369
    GPU:
    KFA2 4070Ti EXG.v2
    Yea truth is I'm not too enthusiastic about system maintenance either, feels like a chore.
    Hope the voltage increases help! If not, we'll work out something else.
     
  15. flow

    flow Maha Guru

    Messages:
    1,023
    Likes Received:
    17
    GPU:
    Asus TUF RTX3080 oc
    I had this in the past. Sudden crash to desktop and/or spontaneous reboots.
    The pc build was a year old. So I started troubleshooting with the ram, and sure enough it didn't sustained the timings anymore. I lowered them and got stability back.
    However, prior to that I noticed an increase in crashes and reboots. Starts subtle, and after some days it happens more and more. So I swapped out the ram for new ones to be sure.

    You should run memtest86+, it takes some time but will stress your ram and find errors if they are present.
     

  16. shimyns

    shimyns Guest

    Messages:
    1,665
    Likes Received:
    1
    GPU:
    SAPPHIRE r9 290x Tri-X OC
    Well guys, my PC's been completely stable since my last post. My CPU is at its normal overclock but my RAM multiplier has been reduced, resulting in 1200 Mhz instead of 1600.

    So, I'm positive it is memory-related. I considered playing with the RAM timings, playing with the DRAM voltage and QPI voltage, too. I started looking into the various functions of the confusing nomenclature of QPI, VTT/QPI, CPU PLL, QPI PLL and then was like screw this. I'm honestly fine with my RAM at 1200 Mhz if that means I can keep my CPU OC and my system is stable, and I just don't want to spend any time squeezing out those extra 400 Mhz (which probably do not affect my performance to any reasonable degree).

    I'm glad it's stable now and all parts seem fine (especially my GPU!). Thanks for your suggestions and if I do end up revisiting this issue, I'll be sure to post back.
     
  17. paka1

    paka1 Guest

    Messages:
    5
    Likes Received:
    0
    GPU:
    i7
    No problem with my Asus x99 deluxe.
     

Share This Page