- May 19, 2011
- 17,574
- 9,261
- 136
I have a customer's computer in at the moment because while a problem with it looks like it has all the hallmarks of faulty memory, it may not be quite as simple as that.
The symptom they reported is that once the computer has been running for about a couple of hours, it might reboot, BSOD, freeze, etc.
Running BlueScreenView to get BSOD summaries strongly suggests faulty memory, including MEMORY_MANAGEMENT, PFN_LIST_CORRUPT, etc.
Here it took somewhat more than a couple of hours for the faulty memory symptom to show itself: I ran two passes with memtest86+ 5.01, then ran prime95 for 45 minutes, (no problems reported so far) then I wondered whether there might be a chance as this machine doesn't really have enough RAM (2GB, 64-bit Windows), I wonder whether a HDD issue could show up as a 'memory issue' due to heavy use of virtual memory. I got the machine going on a full disk check. I watched it finish and reboot, then the machine started acting up (within seconds of Windows attempting to boot, the machine rebooted and then wanted to do startup repair, same again if I chose normal boot). I then ran memtest86+ immediately and got thousands of memory errors.
I replaced the RAM quickly and started memtest86+ for a couple of minutes (I didn't bother to do a full cycle at this point because if it didn't complain straight away then chances are the issue wouldn't show itself in the first 60 minutes based on the evidence). The chkdsk log shows nothing out of the ordinary, SMART readings are fine. As the previous appointment was malware related (the memory issue has been going on for years though), I've got it running a full malwarebytes scan. Next I'll probably leave it running memtest86 overnight.
The question from the title is based on me wondering whether the memory might just need reseating. Normally for simple faulty RAM situations (eg. no POST) I'd try this quickly anyway, but TBH I can't remember the last time that a reseat actually helped and I was wondering what other people thought about it (apart from "it's surely worth a try?"). I'll probably give it a try if I get the chance.
The symptom they reported is that once the computer has been running for about a couple of hours, it might reboot, BSOD, freeze, etc.
Running BlueScreenView to get BSOD summaries strongly suggests faulty memory, including MEMORY_MANAGEMENT, PFN_LIST_CORRUPT, etc.
Here it took somewhat more than a couple of hours for the faulty memory symptom to show itself: I ran two passes with memtest86+ 5.01, then ran prime95 for 45 minutes, (no problems reported so far) then I wondered whether there might be a chance as this machine doesn't really have enough RAM (2GB, 64-bit Windows), I wonder whether a HDD issue could show up as a 'memory issue' due to heavy use of virtual memory. I got the machine going on a full disk check. I watched it finish and reboot, then the machine started acting up (within seconds of Windows attempting to boot, the machine rebooted and then wanted to do startup repair, same again if I chose normal boot). I then ran memtest86+ immediately and got thousands of memory errors.
I replaced the RAM quickly and started memtest86+ for a couple of minutes (I didn't bother to do a full cycle at this point because if it didn't complain straight away then chances are the issue wouldn't show itself in the first 60 minutes based on the evidence). The chkdsk log shows nothing out of the ordinary, SMART readings are fine. As the previous appointment was malware related (the memory issue has been going on for years though), I've got it running a full malwarebytes scan. Next I'll probably leave it running memtest86 overnight.
The question from the title is based on me wondering whether the memory might just need reseating. Normally for simple faulty RAM situations (eg. no POST) I'd try this quickly anyway, but TBH I can't remember the last time that a reseat actually helped and I was wondering what other people thought about it (apart from "it's surely worth a try?"). I'll probably give it a try if I get the chance.