Closed Solved

Need help diagnose/fix possible memory problem

I've been getting intermittant freezes and BSODs on this computer:

Gigabyte P55-UD2
i5 750 cpu
4 GB G Skill DDR3 1600 memory running at 1333
NVIDIA GeForce GT 220
Windows 7 64-bit

Windows has recently suggested running their memory diagnostic. I had already done that, suspecting memory problems anyway. I've run it five times total. It reported a problem once. I've also run Memtest86+ 4.20 five times; twice I let it run for seven passes. It never reported a problem.

Is there anything else I can do to pin down this down? I don't want to spend $50 on new memory if it's not going to solve the problem.

16 answers Last reply Best Answer
More about need diagnose memory problem
  1. start with the simple stuff first...check that you have your mb bios updates...sometime they add new cpu code and bug fixes. run your hd vendor drive tools and look at the drive smart info. use cpu-z check that the mb is reading your ram right and set the timing of the ram right. check that the ram and all cards and power plugs are seated. run ccleaner and defrag..malware bytes and a good anti virus scan...try running prime85 with blend turned on see if thta can force a hang. alos use hardware monitor to watch your cpu and gpu temps and the ps voltage.
  2. thanks, smorizio, for the quick reply. Here's the status:

    - bios version is f11; that's the latest
    - already ran the HD diagnostics; no errors reported
    - HD SMART info all OK
    - already checked CPU-Z info and the settings for timing, etc, agree with bios and with what G Skill and Gigabyte say to use. (Although I could never find a really complete list for this specific MB)
    - I run CCleaner every night and Malwarebytes once a week
    - temperatures are OK; cpu below 50 C (usually below 45), graphics below 55 C (usually below 50)

    I'll try the prime number thing (Prime95?) but even if it does produce a hang, will that tell me anything diagnostic? The blue screens haven't been consistent--different processes, different errors. The logs also have a good number of critical and error events but I'm not informed enough to make anything out of them.
  3. How often do these freezes happen? What are you doing when they happen?
  4. Raiddinn said:
    How often do these freezes happen? What are you doing when they happen?

    Sorry for the slow reply. I guess the email notification didn't work; I didn't realize you had posted.

    I can't find any consistent pattern to the failures. The most common occurs just from starting the computer. It boots, starts Windows, logs me on and looks fine. Then, without my doing anything, within a few minutes it either freezes or throws a blue screen. These are some others I recall: twice it failed starting either a manual update or scan with MS Security Essentials. Twice it froze starting a scheduled recording of a TV program from Windows Media Center. Others, as far as I can remember, seemed random, IE9, etc. I can start making notes to get better information.

    This computer is used mainly as a media center and file server. I should mention about the TV recording. The device is a Hauppauge WinTV-HVR 1850. It has a Microsoft certified driver but I thought it might be suspicious anyway. I removed the board and driver but still had system failures. I believe it's innocent.

    I have several crash dumps that could conceivably be helpful, but I dread the idea of trying to tackle them. I guess I may have to.

    Thanks for your reply. Let me know if there's other things I can do.

  5. Sounds like a flaw in a core component to me. Something physically wrong with the processor, the motherboard, or the RAM.

    Bad driver and bad PSU are also possibilities. Those would apply a whole lot more if you were putting stress on the computer, though, like if you were gaming when things happened. IE9 doesn't use a lot of graphics resources or power so both of these are kinda unlikely from where I sit. Still on the table, though.

    Can you borrow replacement parts for things in your core? Do you know anyone with similar RAM, a similar processor, a similar motherboard?

    You could take your parts to them or them take their parts to you, either way would work.

    You can also download a program called Memtest86+ and use a program like ImgBurn to deconstruct the ISO into individual files and copy them to the CD, then boot up with the CD in the drive and 1 RAM stick in the computer.

    The RAM sticks need to be tested individually with this program, and the program needs to get through at least 7 whole passes per stick if you are going to attempt this (which I would suggest at this point since it doesnt cost $ or rely on other people to give you hardware).
  6. I don't have anyone handy to swap with. Wish I did.

    Windows already suggested running Memory Diagnostic. I ran their utility five times and it reported and error once. Also have run Memtest86+ 4.2 five times. Two of those times I ran for 7 passes. It did not report any errors. My original question was actually whether that one positive from the Windows diagnostic justified replacing the memory.

    The other suspicion I had (besides the TV card) was the PSU. The MB has an 8-pin CPU power receptacle and the Antec PSU has a 4-pin plug. So I used an adapter to extend to 8 pins. This didn't seem like a problem because the computer really loafs nearly all the time. Also, before I installed Windows, I ran Ubuntu for about 7 months and it never had any problems.
  7. sounds like one of the ram sticks has a bad bit. i would see if the vendor can do an express cross ship so you dont have to pull the ram till it get into your hands.
  8. When you ran those memory tests did you have 1 stick in the computer that you rotated between tests?

    The Windows version is the less reliable of the two since it can't touch any RAM that is reserved for the OS, so it just never checks huge swaths of the usable RAM.

    However, these programs need to be run on 1 stick of RAM at a time, because they can only test a small block at one time and if you have lots of RAM inserted you have no idea where that block is going to end up.
  9. Didn't know that. Thanks. I'll run Memtest again on each stick separately.

    What about the 4-pin, 8-pin CPU power plug question? Do you think that's likely to be an issue?
  10. It could be, and it would be easy for me to blame everything on that since its clearly not what was intended to be used and stuff, but my gut feeling says that's probably not it.

    It may be the case that the PSU just isn't powerful enough to handle the system anymore. Even if it was working in the past the components in it degrade through usage so something that performed well for a few years is not guaranteed to perform well for a few more.

    The cables on those wires for the most part aren't super overloaded anyway. The AMD processors are often 125w or more and they use 4 pin sockets with 2 grounds so that means each 12v wire on the processor power cable is doing more than 60w.

    The Intel processors usually pull less than that.

    It is a possibility, but it doesn't seem like a strong one.

    What is your current PSU and how old is it?
  11. Sorry again, I was out of town and pretty much out of touch. I've now run the Memtest86+ tests on individual sticks and didn't get any reported errors.

    I've made sure my drivers are up to date. They already were except there was a more recent one for the TV card that I have now installed. No real reason to think it will help since I had failures when that card wasn't even installed.

    I do have one new piece of information. I had already run a test on the HD using the Seagate tools. It found nothing even with the long test. But I tried testing again with the Western Digital diagnostics (drive is a WDC WD640). That test, DLGDIAG 5.04f, didn't really even start. It reported a DRQ timeout, error/status code 0135. Looked that up and here's what they say:

    Timeout from checking Data ReQuest Timeout (DRQ) bit. The drive has not responded back in the time allotted. This may be due to a defect with the drive or a bad connection. Check cable & retest. Replace the drive if the error repeats.

    I replaced the cable, vacuumed the connectors, and ran it again with the same result. Because of the case and board design I can't try connecting the drive to a number of different board sata connectors as I'd like to; the case would need major hacksaw modifications. I may do that anyway; I'm getting frustrated with this problem.

    I guess the error message is clear enough but it seems funny that neither Seagate nor Windows found a problem.

    Power supply in an Antec Earthwatts 380 Watts. Bought it at the same time as the rest of the components, early 2010.

    Thanks for your replies. I really appreciate your help.
  12. Best answer
    I trust the Antec to power that setup even after 2 years of usage.

    Try downloading AIDA64 and using it to crush the processor.

    If that doesn't fail, there are other tests you can do with it too and you might want to try those.

    Do you have the potential to borrow a different PSU from somebody?

    How about a different hard drive?
  13. Best answer selected by arby10101.
  14. I think I have come up with an OK way to swap a known-good hard drive. I'll do that this weekend. (It's too hot here to feel like going out much anyway.) I'll run that for a while and see whether it helps.

    I'm going to mark your last post as the answer. I don't think we're going to find this without being able to do a lot of parts swapping. I plan to replace my work computer after the summer and I'll have parts then to check everything. Until then, if the hard drive doesn't help I can put up with the crashes; they're intermittent.

    Thanks, Raiddinn, for your help with this and your patience.
  15. Parts swapping is usually the best way to figure out any problem, sadly its not easy for most people to accomplish. Probably about half the people that come here have at least some parts laying around to switch out and only something like 10% of all the people happen to have laying around what I need them to try switching out.

    Usually we must resort to testing in non-conclusive ways and using guesswork.
  16. This topic has been closed by Nikorr
Ask a new question

Read More

G.SKILL Memory