• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Memory BSOD?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

ControlledChaos

Registered
Joined
Jun 2, 2011
Location
Weymouth
First I want to say hello as this is my first post but I have been learning and lurking here for a while now.

So my issue is I have been getting a few BSOD the past few days. They seem totally random at times. The message that has popped up more than once is Memory Management error. So this lead me to run Memtest86 on my machine thinking that I might have a bad stick of RAM but it passed. The next night I'm doing some light gaming and BAM BSOD again with the same message. So again I run Memtest and it passes again. Now I just want to get to the bottom of this as crashing in the middle of something is annoying as hell. So i break out the old Prime95 and run it. 5-10 minutes in all 4 workers stop so now I'm really befungled. I decide to run Memtest one more time and finally get a single error. After that it got late and I went to sleep.

So that brings me to today at work I go through the process of testing each stick alone and testing each mem slot to see if I can narrow it down and get nothing. Everything passed wtf?! Put all the RAM back into my machine and passed as well. So I run Prime 95 again and 2 of the workers fail so I hit it with Memtest again and again just a single error pops up. Reboot after the test and run it once more and it passed.

So it seems only after running a blend test with Prime95 and getting an error does my RAM error on a Memtest as well. I'm lost as to what it could be. I have just recently updated my BIOS and loaded the default settings so maybe I'm wondering if I have to change a setting or something or maybe my board is wacked. The system has been running fine for the past 3 months and only started this recently.

basic run down of the system is I5 2500k with a Gigabyte P67A-UD4-B3 board with 12 Gigs of Kingston HyperX DDR3. 2 OCZ Agility 2 in Raid 0 and a 1TB WD black as storage and PNY GTX560 OC2 in SLi with a 750watt OCZ Fatility PSU in a Thermaltake Element Nvidia edition case.

Air flow seems good in the case and I tested the temp of RAM as I ran Prime95 and they were about 100 degrees f at load in the blend test of Prime95. I have a new 850 watt PSU on its way as i didnt like the OCZ one anyways but hoping that might be my issue but figured I would post here and ask as well for some advice.
 
It is the RAM. I had a bad stick that tested perfectly in memtest86+ but gave me BSoDs. It was worse when I was using both sticks together for some reason, but only one stick was bad. The good stick alone ran my computer fine, but the bad stick made programs randomly crash for about 15 minutes before a BSoD. Both sticks together was an instant BSoD as soon as Windows loaded. The memtest threw me off for a few days because I thought it was fine, but I RMA'd and got a replacement pair, and my system has been working fine ever since.
 
The problem with that is which stick is it? How do I go about RMAing it? I have a 4Gig kit and an 8 gig kit. I have tested each stick individually and they all passed so there is no way to single out the bad RAM. I have also switched up how I had them installed in the slots. 4x2x4x2 now in each of my slots and haven't had a problem yet. I have been experiencing some slight stuttering in one of my games but I'm not sure if that's the Logitech app running in the back ground or not causing that.
 
Wait wait wait wait wait...

Wait...

A 4GB kit (as in 2GBx2) and an 8GB kit (as in 4GBx2)?

Also, what brands are they?
 
Wakirimas Controlled Chaos,

This probably is not the case for you but at this point you should consider everything and anything. I had symptoms similar to yours on one of my rigs years ago. I had it built and bought it from a local shop. I had them install extra memory after about a year to upgrade the unit.

In my case the RAM was grounding out against the case (it was a micro). It was a tight fit and when things warmed up it was enough expansion to cause it to ground out. When I removed the case sides and exposed the RAM it never happened. I made a case modification and problem solved. I had it back in the shop numerous times and they never found it. I discovered it through trial and error. I'm still surprised I found it. :eek:

If it is bad RAM my only other thought on tracking it down is via Windows event viewer. This is something a Windows guru might be better equipped to answer but I believe memory faults are often recorded with a memory location the fault occured at. I would imagine that the faulting address would tell you which bank caused the fault and thus which stick. This is a guess on my part but it might be of some value.

Regarding memtest, I am assuming that it may not be perfect as Aldakoopa suggests. Running those tests are pretty boring and I'm not sure they really give you the wide range of enviroments you'd find in memory usage. I would suggest running them as matched pairs and run various games and programs. The culprit should make itself known. Since they are kits you'll have to RMA the entire kit since they are matched (I assume). I had to do that with my Corsair when I built this rig.

Also remember that only sticks that are tested together are "guaranteed". As I understand it two kits should work together but are not guaranteed to do so because they aren't physically testing all of the sticks together at the same time. So guaranteed kit plus guaranteed kit equals should work but might not.

Hope those shots in the dark help.

Sincerely,

Old School Ninja :sn:
 
Well the machine has been running fine all night but I haven't really tried any heavy gaming on it.

I was looking at the heat on the RAM but im not sure what an acceptable temp for RAM is under load. I figure around 100 degrees F is pretty normal.

I have my new PSU on its way and plan on doing a full reinstall of the OS tomorrow so that should squash any of those bugs that may have been there.

Im not to experienced with Prime95 and knowing what the results mean when it fails so maybe I will have to look into that more.
 
Use Intel Burn Test. Forget Prime95 for a while.

Run 5 to 10 passes on standard.

If you may, post screenshots here.
 
Well I got my new PSU from the UPS guy today which also came with a 4 Gig Kingston HyperX DDR3 stick for free (Thanks NewEgg combo deal) I swapped out one of the 2 Gig sticks in my machine for the new 4 Gig to bring the total to 14 Gigs. I installed my new PSU and reinstalled my OS on a single SSD and used the other for games I want to load faster.

No issues so far but I haven't tried to stress test it or anything. I did run a Memtest after installing the new RAM and it passed as I knew it would. The real test is after I get everything up and running and doing some gaming on it to see if the issue still rears its ugly head and then I know I didn't fix it.
 
Back