memtest errors and then no more error

K

kandor

Gawd
Joined
Aug 24, 2005
Messages
632
Hi,

Suddenly my system bsod'd 4 times in 24 hours. I ran a memtest, immediately saw errors, let it finish over night, something like 900k errors by morning. (8x4gb sticks).

I run windows for a bit, reboot memtest, no more errors (at least for an hour, then I removed the ram, needed to work).

Now I testing those sticks in another system, no errors there either. So, does this point to a mobo failure?

Now the system is using on 2 sticks (8gb each) and is stable all day. Seems really strange.

These are the 4 bsod's:
071415-8782-01.dmp 7/14/2015 11:08:09 PM MEMORY_MANAGEMENT 0x0000001a 00000000`00041790 fffffa80`0bd93f00 00000000`0000ffff 00000000`00000000 ntoskrnl.exe ntoskrnl.exe
071415-8642-01.dmp 7/14/2015 9:54:28 AM DRIVER_IRQL_NOT_LESS_OR_EQUAL 0x000000d1 ffffd800`036890aa 00000000`00000002 00000000`00000008 ffffd800`036890aa ntoskrnl.exe ntoskrnl.exe
071415-9282-01.dmp 7/14/2015 2:23:07 AM PAGE_FAULT_IN_NONPAGED_AREA 0x00000050 ffffdca0`27021098 00000000`00000000 fffff880`01c5872f 00000000`00000007 ntoskrnl.exe ntoskrnl.exe
071315-11216-01.dmp 7/13/2015 10:21:17 PM KMODE_EXCEPTION_NOT_HANDLED 0x0000001e ffffffff`c0000005 fffff800`034774b4 00000000`00000000 ffffffff`ffffffff ntoskrnl.exe ntoskrnl.exe

system spec:
i7-4930k @ 3.4 Asus Sabertooth X79 - 32 GB ram, quadro 4000k
 
hci memtest has worked the best for me

have tried
memtest86 old version (non-uefi)
memtest86 uefi
memtest86+
gold memory tester

Used hci recently to try to get 32gb to work in a z87 @ 1866 which stressed the memory controller and was a pain. Found errors that the other programs didnt.
 
Hi, using memtest UEFI 6.1.0 I found the following.

The computer had 2 sets of ram, i.e. 2 separate kits, 4x4gb. Both are corsair vengeance, but were purchased at different times.

I took all of them out of the machine. I tested the first kit in another computer overnight, no errors.

The second kit tested as follows (note I tested these 1 at a time).

First stick, 2 errors
Second stick, 8 errors
Third stick, 1,411 334 errors and counting
Fourth stick, not done yet.

Seems weird that a whole kit would fail like that?
 
If the memory tests okay elsewhere, I would begin to question the memory controller on the cpu. And recommend hci memtest over anything.

I'm not super familiar with that chipset/cpu and the limits of, how long was it previously running without problems?

What memory speed are your kits?
 
Hi, kits are 1600mhz 9-9-9-24

All 8 sticks ran fine on that mobo for 6-8 months. The 4 blue sticks run fine. The 4 black sticks all have memtest errors. The problem started afaik on the 13th with the first bsod.
 
kandor said:
Hi, using memtest UEFI 6.1.0 I found the following.

The computer had 2 sets of ram, i.e. 2 separate kits, 4x4gb. Both are corsair vengeance, but were purchased at different times.

I took all of them out of the machine. I tested the first kit in another computer overnight, no errors.

The second kit tested as follows (note I tested these 1 at a time).

First stick, 2 errors
Second stick, 8 errors
Third stick, 1,411 334 errors and counting
Fourth stick, not done yet.

Seems weird that a whole kit would fail like that?
Click to expand...

Possible, I know I've had real bad luck with Corsair... That being said, I would double check the bios auto-settings make sense. Then definitely re-test the sticks in another rig to make sure if it's really the sticks.
 
hmm that is a little weird, im guessing youve tried to test different memory slots? have you tested "good" sticks in one of the same slots the "bad" ones failed? For example, test 1 black in the far right slot, remove; install blue and test again. This might help rule out the quad channel memory controller

If I were you, i'd try posting in the corsair official forums and see what they have to say. They will likely ask you to re-test the same stuff you might have already done

Seems really complicated. I'd double check memory settings in the bios and make sure they are all at stock and 2n.
/edit did you have to mess with the vtt voltage when you set it up the first time?
 
Last edited:
I checked the black sticks one stick at a time in a single slot and they all failed, the blue sticks that came from the same machine (but a different kit) all passed in the same slot. I think I'm ready to RMA these. Thanks for the advice!
 
kandor said:
Hi, kits are 1600mhz 9-9-9-24
Click to expand...
Only if they're made with Nanya "DG" chips, but most likely the chips are 1600 MHz, 11-11-11-xx, just like Nanya DI or Hynix PB or Samsung K0 or Micron -125 -- at best. But almost all 1600 MHz desktop modules sold now are made from 1333 MHz chips or mystery chips, and I've seen 1333 MHz chips in 1600 MHz Corsair Vengeance Pro.
 
HI, these are just printed 9-9-9-24 on the side, Not sure who made the chips.
 
On the new kit, did you try bumping the voltage or lowering the freq and loosening the timings?
 
No, all 8 sticks worked together just fine at stock settings for around 6 months, when the system suddenly started to bsod, and the 4 newer sticks showed memtest errors.
 
You must log in or register to reply here.
Back
Top