Memtest86+ 1.70, 24% = Pass

Discussion in 'OT Technology' started by Tenny, Jun 5, 2008.

  1. Tenny

    Tenny OT Supporter

    Joined:
    Oct 4, 2005
    Messages:
    13,876
    Likes Received:
    0
    Location:
    Bay Area, CA
    System Stats:
    Intel C2D E6750
    Gigabyte GA-P35C-DS3R
    EVGA 8800GT Super Clocked Edition
    2x 1GB Crucial Ballistix Tracer (4-4-4-12)
    --Older style Crucial Ballistix (2x LED Columns)
    2x 1GB Crucial Ballistix Tracer (4-4-4-12)
    --New Style, supposed to be worse. (1x LED Column)


    Background:
    Built this system last September. Ran some OCZ ram for a while, then got a good deal on the Crucial for maybe $40 for both. Computer has been a champ almost the whole time. It has a few issues running Assassin's Creed, but I think that's more of an 8800GT issue as it was very common among several other machines.

    Flash forward to recent:
    So I got my two new sticks of Crucial, put them in the machine. Booted up fine and everything. Gave Memtest86+ a run, got to 24% and then it "passed" me. I was eager to play on it so I was like wtf, whatever and went to play games. After a few hours of playing games I ran into problems. The game (TF2) gave me some error (don't have it available now). Told friend, he said he had the same problem once, and restarted to fix it. No dice for me. So I started Memtest86+ up again, and let it run through the night. I woke up and it had 2 passes, and maybe 5 errors.

    So I knew something was wrong obviously. Once I got back from work today, I took out the old memory, and ran Memtest86+ again. Ran 3 passes (each pass got to 100%) so everything appeared fine, figured maybe my old ram was bad? Took out the new ram, put the old ram back in (all staying to the original slots they were in the previous day) and ran 3 passes just fine (all the way to 100% again...)

    So I am kind of confused on whats happening, and decide to throw the new ram in again, while keeping the old ram in. So far it's passed twice in a row, now just got an error.

    Test: 6 Pass: 2 Failing Address: 0008a8329fc - 2216.1mb good: 00200000 bad: 00000000 err-bits: 00200000 count: 1.


    So I see two problems.
    (Semi-Cliffs)
    1. Why is MemTest86+ 1.70 only getting to 24%
    2. Why is the memory failing before, and not now? (now failing on third pass)
    --I am pretty sure I seated the ram fine, but that's my only idea of why.
     
    Last edited: Jun 5, 2008
  2. DAN513

    DAN513 OT Supporter

    Joined:
    Mar 10, 2003
    Messages:
    10,089
    Likes Received:
    2
    Location:
    204
    if you have any errors, you have bad ram. Replace the faulty pair.
     
  3. Tenny

    Tenny OT Supporter

    Joined:
    Oct 4, 2005
    Messages:
    13,876
    Likes Received:
    0
    Location:
    Bay Area, CA
    As I stated, when I run either of them by themselves, they both pass fine.
     
  4. Chris

    Chris New Member

    Joined:
    Oct 27, 2003
    Messages:
    14,711
    Likes Received:
    0
    Location:
    Texas on my mind
    possibly motherboard, have you tried single sticks in each of the slots?
     
  5. Tenny

    Tenny OT Supporter

    Joined:
    Oct 4, 2005
    Messages:
    13,876
    Likes Received:
    0
    Location:
    Bay Area, CA
    Yes.

    The old memory is in the same slots it was tested in alone, along with the new in the same slots it was tested in alone.

    I just upgraded to the newest Memtest86+ (2.01), and ran Test #6 and it passed fine. I am wondering if it was a problem with Memtest86+ 1.70, as right now I am at 25%. I'll let this run for a few hours before reporting back...
     
  6. Tenny

    Tenny OT Supporter

    Joined:
    Oct 4, 2005
    Messages:
    13,876
    Likes Received:
    0
    Location:
    Bay Area, CA
    Still had errors.

    Both sets of ram do well by themselves.

    Planning on going to pick up some more ram tomorrow 4x1gb or 4x2gb and see if I have errors there...
     
  7. Mike99TA

    Mike99TA I don't have anything clever to put here right now

    Joined:
    Oct 3, 2001
    Messages:
    4,553
    Likes Received:
    0
    Location:
    Greenville, SC
    Try memtest86 3.x (not memtest86+ - they work differently). I've had similar issues with memtest86+ - never had the problem with plain memtest86.
     
  8. Tenny

    Tenny OT Supporter

    Joined:
    Oct 4, 2005
    Messages:
    13,876
    Likes Received:
    0
    Location:
    Bay Area, CA
    Tried Memtest86, same thing.

    It always passes the first pass, then the second it errors. IT doesn't matter if it's been running for 5 hours, then i restart the test, it always passes the first one, then fails the following tests.

    Didn't get new ram, didn't have a chance... will try to over the next few days though...
     
  9. Doomsday

    Doomsday XXX

    Joined:
    Mar 14, 2000
    Messages:
    14,902
    Likes Received:
    0
    Location:
    Minnesota
    is the pc overclocked?
    do you have the right memory voltage?

    stop worrying why it passed on the first pass, and not on the 2nd.
    memtest86 is a ram test tool, not a diagnostic tool.

    if it fails, it's really simple,... something is wrong.

    any RAM failing memtest should be replaced.
    if a known good working ram failed memtest, blame the mobo.
     
  10. Tenny

    Tenny OT Supporter

    Joined:
    Oct 4, 2005
    Messages:
    13,876
    Likes Received:
    0
    Location:
    Bay Area, CA
    Nothing is overclocked. The voltage is the standard voltage.
     
  11. awns729

    awns729 New Member

    Joined:
    Jan 4, 2008
    Messages:
    353
    Likes Received:
    0
    did you double check the ram operating voltage w/ the default voltage your mobo supplies?\

    w/ my first build, my OCZ ram ran at 2.1v, the mobo supplied 1.8. it ran fine (booted windows, gamed okay), and i wouldnt have known had i not had other (probably, IIRC) unrelated errors.
     
  12. Ameter

    Ameter Active Member

    Joined:
    Mar 31, 2005
    Messages:
    97,795
    Likes Received:
    1
    Location:
    Calgary, AB
    do you have the new pairs side by side or spaced? (aka, are they in the proper slots for dual-channel or not?)

    Is one set ECC while the other isn't?
     
  13. Ameter

    Ameter Active Member

    Joined:
    Mar 31, 2005
    Messages:
    97,795
    Likes Received:
    1
    Location:
    Calgary, AB
    are the ram chips the same speed, same CAS latency, etc?
     
  14. P07r0457

    P07r0457 New Member

    Joined:
    Sep 20, 2004
    Messages:
    28,491
    Likes Received:
    0
    Location:
    Southern Oregon
    just because the ram is okay on it's own does not mean it will play nice. Everything has tolerances. It is entirely possible, and actually quite common for those tolerances to "stack" and produce an unstable environment where everything works individually, but not as a whole.


    Think about it this way... You want to measure out one mile. Everyone has a tape measure and tries to measure 3 feet, each. Due to margin of error, lets say they could be +/- 1" on either side, so that's +/- 2" margin of error, for each person. Given that you'd need 1760 people to measure a mile, and each person had a margin of error of 2", you have a cumulative margin of error of 3,520" or 293 feet. Basically you could only reasonably estimate a mile to within a football field. So in the grand scheme of things, a small 2" tolerance all of the sudden becomes a football field!

    Same thing with electronic components -- except tolerances are much much tighter.
     

Share This Page