[RAM] is my RAM bad or is memtest lying to me

[Awsum McPossum]

Hey Guys!

Specs of my PC:
Mobo: ASUS M4A89GTD PRO
CPU: AMD Phenom II x4 970 Black Edition
RAM: G.SKILL Ripjaws Series 8GB (2 x 4GB) 240-Pin DDR3 SDRAM DDR3 1333 (PC3 10666)
VIDEO: ASUS HD Radeon 6870 1GB
CPU COOLER: Xigmatek Gaia SD1283, NOT STOCK
POWER SUPPLY: CORSAIR CMPSU-750TX 750W ATX12V v2.2
HARD DRIVE: Western Digital Caviar Black WD1002FAEX 1TB 7200 RPM 64MB Cache and a second 500GB one for storage
OS: Windows 7 Professional
CASE: Cooler Master HAF 922 Mid Tower

So around September 7th I had a PSU failure and RMA'd a replacement. Got everything up and running again but have been having intermittent bsod crashes since, most often while playing games (The errors are usually either Driver_IRQL errors or memory management errors) and the occasional refusal to post (as an aside, my ASUS mobo has a mem_ok! led that would remain illuminated when this happened). I have checked and updated my drivers multiple times to no avail, ran /chkdsk to see if bad sectors in my hdd were the problem, and ran several system repairs to see if my windows install was to blame. So after reading several other posts I decided that it would probably be a good idea to test my RAM.

So last night I set up and ran memtest with both sticks. Woke up this morning, checked the computer, and memtest was reporting 0 passes and a ton of errors. So I moved on to testing individual sticks (here I'll refer to them as Stick A and Stick B, and the mobo slots as 1,2,3,4). I left Stick B in slot 4 and ran memtest. Immediately it reported errors. So I removed stick B and put Stick A in slot 2. Ran it for 1 pass and it reported no errors. In order to eliminate the slots being fried I put Stick A in slot 4 and ran it for 2 passes with no errors (so I'm assuming now that it is not the RAM slots on my mobo). What has me stumped is I am currently running Stick B in slot 2 and so far it has not reported an error (its on pass 2). So my question to you guys: is my RAM faulty? or is it something else? Thanks for any help!

and yes I know 3-4 passes is the recommended amount, I'll update this post as I let my suspected faulty stick run.

TL;DR RAM fails first memtest run but passes next one. wut?

 

[cl-scott]

What you need to do is test both memory modules in every slot. If you repeatedly get errors on one slot with either module, but not the others, then you would have a very strong case for a bad slot.

 

[JDFan]

Could also be a bit off Vdroop on the MOBO slots when both modules are inserted combined with a module that is barely remaining stable -- try increasing the voltage a single step and rerun the test with both modules installed.

 

[Awsum McPossum]

I'll try upping the voltage. I have been testing them in the other slots. The stick I thought was faulty checked out in all the other ones, and the stick I haven't gotten any errors from checked out ok after 4 passes on the potentially troublesome slot.

Currently running them simultaneously again with stick A in slot 4 and stick B in slot 2. No errors so far.

Just another thing to add, I can't actually use slot 1 since the after market CPU cooler I have sticks out just enough that the RAM with its cooling fins don't fit under the fan.


Also cl-Scott since all the slots have so far checked out, should I just get new RAM?

 

[cl-scott]

It could have just been one of those strange flukes. One of the gods was bored, decided to create a problem for you that you'll spend hours or days chasing, when it was just a random thing. Keep testing, and if you can't duplicate the error, then don't worry about it.

 

[Awsum McPossum]

Well some more frustration.
Upped the voltage and tested out the sticks together. Passed 3 times. Went in to windows and decided to see if I could reproduce the blue screen error. Ran a future mark 11 benchmark and it crashed to a blue screen citing memory management. Both sticks again failed to pass memtest after that so I decided to go ahead and replace the RAM. Got a corsair ddr3 8gb (2x 4gb) kit. Put it in my system and again ran the future mark benchmark. To my dismay it again crashed (this time with no bsod). Just finished running memtest on the new ram and it completed one pass with 53 errors on test 8. I have no idea what I should do now. Any advice? Is my new ram bad? Or is RAM not the problem?

Edit: scratch that. This new ram crash was actually my own fault. Using the memok! Switch on my motherboard when the old ram was installed manually set some timings which are not compatible with the one I just purchased that I wasn't aware of. /facepalm. Fixed the timings and it passed 3d mark 11 A-ok! I'm going to run memtest overnight and I'll let you guys know of any problems, but fingers crossed and knock on wood I think it's fixed. Thanks so much!

 

[Awsum McPossum]

I did have one crash last night while playing BF3. It was a Driver_IRQL crash but memtest passed without any errors. I think RAM was one of the problems but there must be some other underlying problem going on. Since installing the new ram my rig actually feels like its running faster than it has for a while so I think the RAM was at least one of the proximate problems, if not the ultimate. Anyways I guess we can mark this solved for now, and if I have any further blue screens I'll post to that forum.

 

[Awsum McPossum]

Best answer selected by Awsum McPossum.

 

[nayupov]

How ever strange it may seem, but my findings about successfully using 4 sticks of memory modules simultaneously are the following: I MANAGED TO USE ALL 4 STICK ONLY WITH A MEMORY COOLER - in my case I put as a test, one of 120mm case cooler fans very close to memory modules (ripjaws 4x4Gb - 16Gb) and passed 7 hours test in memtest86. Next, disconnected fan and start test again - failed on test 5. Re-ran test with fans on for another 8 hours - brilliant results! Disconnected fans - failed again on 6th test. So on - I tried both ways several times, and successful tests were at least of 7 hours. Try it out guys! Maybe your problem will also be solved this way!
Brgds, Nodir

 

[Awsum McPossum]

So I noticed that a lot of people are viewing this thread and figured I would give a final update. Turns out the dying PSU damaged my mobo when it went, specifically the RAM sockets. The errors were intermittent for a couple months after I posted this, and then got to the point where I couldn't use the computer reliably. Replacing the Mobo and CPU fixed my issues.