Weird RAM errors.

Discussion in 'General Hardware' started by xankazo, May 10, 2014.

  1. xankazo

    xankazo Ancient Guru

    Messages:
    2,811
    Likes Received:
    1
    GPU:
    EVGA GTX 1060 SC (6GB)
    I got an error message on Event Viewer (I'm one of 'em Event Viewer paranoids) suggesting to check for memory errors. So I loaded up Memtest86 on an USB Flash disk and effectively one single error popped up after 4 passes. I repeated the test 3 more times and every time the same error came up.

    Why I'm saying this is weird? Because my system runs and behave totally stable, I do everything including gaming and using VMWare and there's not a single crash or even lockup or slowdown. Nothing.

    I ran down bad RAMs before and it always meant a lousy experience. So, what's the catch here? I'm was planning to upgrade to 16GB on July anyways, but now I'm not sure if the faulty RAM will get there or not. *confused*
     
    Last edited: May 10, 2014
  2. IcE

    IcE Don Snow Staff Member

    Messages:
    10,659
    Likes Received:
    12
    GPU:
    Gigabyte 970 G1
    Even 1 error detected means hardware failure. Find the defective stick and replace it. The errors can only get worse.
     
  3. clawhamer

    clawhamer Ancient Guru

    Messages:
    2,579
    Likes Received:
    0
    GPU:
    EVGA 780 Classified
    Maybe give the ram voltage a bump or two ?
     
  4. xankazo

    xankazo Ancient Guru

    Messages:
    2,811
    Likes Received:
    1
    GPU:
    EVGA GTX 1060 SC (6GB)
    OK. It's at 1.5v with all stock speeds and XMP selected. You think 1.6v should be fine?
     

  5. clawhamer

    clawhamer Ancient Guru

    Messages:
    2,579
    Likes Received:
    0
    GPU:
    EVGA 780 Classified
    I wouldn't go quite that high to start with, just give the plus key a couple of taps.
     
  6. AsiJu

    AsiJu Ancient Guru

    Messages:
    3,659
    Likes Received:
    18
    GPU:
    EVGA 980Ti SC+
    Try disabling XMP and rerun the tests.

    If it works, update your mobo BIOS and try re-renabling the XMP profile.

    edit: upgrade your BIOS anyway if you haven't yet. The Asus Z87 series has received numerous updates which enhance stability (the Maximus VI has the Z87 chipset).
     
    Last edited: May 11, 2014
  7. Corrupt^

    Corrupt^ Ancient Guru

    Messages:
    6,597
    Likes Received:
    1
    GPU:
    ASUS 1080GTX STRIX
    Yep.

    On a stable system memtest should be able to run for hours on end with cores set to parallel without a single error.
     
  8. sykozis

    sykozis Ancient Guru

    Messages:
    20,017
    Likes Received:
    30
    GPU:
    XFX RX 470
    If you have the option to do so, set memory voltage to 1.55v and retest.
     
  9. xankazo

    xankazo Ancient Guru

    Messages:
    2,811
    Likes Received:
    1
    GPU:
    EVGA GTX 1060 SC (6GB)
    Alright, this is getting even weirder. I ran the test again. I let it finish 3 passes and no errors came up. Then ran it again and that single 1-bit error happened in pass 4. It has always happened in test 8 in the same spot (4927MB).

    I noticed memetest86 was only using 1 core so I enabled all cores in parallel. Since all errors happen in test 8, I ran only test 8 for 5 passes and no errors. Next step would be to let it run overnight with all cores enabled in parallel and see.

    All this time and not a single crash, slowdown, hiccup has happened while using Windows 8.1 Pro. I'm considering to forget about this and enjoy my build.
     
  10. sykozis

    sykozis Ancient Guru

    Messages:
    20,017
    Likes Received:
    30
    GPU:
    XFX RX 470
    more passes = more heat = more likely to generate errors.

    Let it run over night and see if the error repeats consistently.
     

  11. xankazo

    xankazo Ancient Guru

    Messages:
    2,811
    Likes Received:
    1
    GPU:
    EVGA GTX 1060 SC (6GB)
    It makes sense, but once it gets to that point, shouldn't it keep happening? It hasn't been the case. The error has always and exclusively happened in pass 4, but pass 5 comes up clean.
     
  12. sykozis

    sykozis Ancient Guru

    Messages:
    20,017
    Likes Received:
    30
    GPU:
    XFX RX 470
    Let memtest run over night. If there's no consistency to the error, just ignore it.
     
  13. xankazo

    xankazo Ancient Guru

    Messages:
    2,811
    Likes Received:
    1
    GPU:
    EVGA GTX 1060 SC (6GB)
    Alright. I'll post back the results. Thanks.
     
  14. xankazo

    xankazo Ancient Guru

    Messages:
    2,811
    Likes Received:
    1
    GPU:
    EVGA GTX 1060 SC (6GB)
    Well, I ran the test all night for 12 passes and not a single bleep. Here's what I changed: I upped the RAM voltage two steps and enabled all cores (parallel) in the test. Here's a summary:

    [​IMG]
     
  15. clawhamer

    clawhamer Ancient Guru

    Messages:
    2,579
    Likes Received:
    0
    GPU:
    EVGA 780 Classified
    Well, that's sounding positive, now just have to keep an eye on Event Viewer to see of the error message stops.
     

  16. xankazo

    xankazo Ancient Guru

    Messages:
    2,811
    Likes Received:
    1
    GPU:
    EVGA GTX 1060 SC (6GB)
    I always keep checking Event Viewer and the warning only appeared once last Thursday and never again. I guess I'll just chill out and use my rig with confidence. Anyways I'll be upgrading the CPU heatsink because the stock one is utter crap.

    Thanks guys for all the tips.
     
  17. Corrupt^

    Corrupt^ Ancient Guru

    Messages:
    6,597
    Likes Received:
    1
    GPU:
    ASUS 1080GTX STRIX
    Good.

    The parallel is just to make sure you won't run into memory errors in multicore apps. I had broken ram before which only had problems in CSGO, VMWare Workstation and BF4.

    They all had 1 thing in common... all 3 applications had multicore support. Not a single error in single core mode but a ton in parallel.
     
  18. xankazo

    xankazo Ancient Guru

    Messages:
    2,811
    Likes Received:
    1
    GPU:
    EVGA GTX 1060 SC (6GB)
    Yeah. Thank you for the tip. I use VMware extensively for long periods of time and not a single blip.
     
  19. xankazo

    xankazo Ancient Guru

    Messages:
    2,811
    Likes Received:
    1
    GPU:
    EVGA GTX 1060 SC (6GB)
    I guess bad RAM is bad RAM.
    My system started crashing and now lots of errors are showing in memory test apps.
    I removed the bad RAM and unfortunately I will have to run 4GB on single channel until I can get a new kit, next month (tight budget).

    So it's true what they say: Even 1 single memory error is bad. RMA'ing is not an option because paying for shipping from the DR is almost as expensive as buying new RAM.
     
  20. Corrupt^

    Corrupt^ Ancient Guru

    Messages:
    6,597
    Likes Received:
    1
    GPU:
    ASUS 1080GTX STRIX
    Not that I wanna bash Corsair but lately me and some colleagues had lots of issues with Corsair RAM, similar to yours.

    When I bought this PC I bought it with 8GB Vengeance 1600 Mhz RAM, it has been rock solid stable, memtest parallel even at 1866 Mhz. 3 years later I buy another set of the EXACT same type of ram (just a newer batch number of course) and I was suffering of issues very similar to you.

    I also have 8 GB of worthless ram lying around here.
     

Share This Page