Hello,
Im having an odd issue with frequent BSOD's on a new system build that I cant seem to narrow down. Below are the system specs:
Motherboard: Gigabyte z68x-ud3h-b3 Revision 1.3
RAM: Mushkin Blackline 996995 2x4GB (4 DIMMS)
CPU: Core I7 2600 (non-k)
GPU: XFX Radeon 6950
PSU: OCZ700MXSP - 700w
OS: Windows 7
Alternate OS: Ubuntu 12.04
Issue:
The problem Im having is that the system isnt stable using 4 DIMMS at default values - 2 kits of the Blackline 996995 2x4GB for a total of 16GB of RAM. If I take 2 DIMMS out I can run stable without an issue, even enable XMP to bump the ram up to 1600. Put all 4 DIMMS back in and boot, and its BSOD and system hangs during every boot with XMP enabled. turn off XMP (back to 1333Mhz) and it still hangs or BSOD but not as frequently.
The only way I can seem to get it running stable is by dropping the RAM speed down to 800Mhz.
Testing notes:
- System BIOS is currently at U1H (UEFI BIOS) and was previously at F10 - issue happened with both bios versions.
- Ran MEMTEST86+ 4.20 on each DIMM for 3 hours each. No errors. This is with XMP enabled.
- Ran MEMTEST86+ 4.20 with all 4 DIMMS installed for 3 hours. still no errors. again, this is with XMP enabled.
As soon as I try to boot into windows: BSOD, or system freeze.
I know memtest isnt the greatest stress tester for a system build, but I just havent experienced a system build where each DIMM passes several hours of testing and yet BSOD's every single time. Could it be CPU related? Memory controller or CPU Cache? I would think if it was, then I would be seeing this regardless of how many DIMMS I have installed...
- tried booting the system with only 3 DIMMS installed and XMP enabled. Still crashes.
- tried booting with 1 DIMM installed and XMP enabled. No issues booting.
Could the memory controller be having issues running in dual channel mode with that high of a memory frequency?
As a suggestion from Mushkin support, I tried changing the QPI/VTT voltage. I tested in increments of 0.010v from 1.075v to 1.125v for the QPI/VTT and still BSOD.
On a whim, I decided to unplug the Radeon and just run with the onboard graphics, since everything else was disconnected exept for the HDD so I could boot into something. Restarted the computer and it boots into windows without a problem. I was shocked, put the 6950 back in, and boom - blue screens again/crashes in Ubuntu.
Im convinced now that its something else, so i decided to perform some different troubleshooting:
- Removed Videocard and ran system with onboard graphics: System appears stable at XMP profile.
- Tried running videocard without PCIe power connected: will not boot. (Duh!)
- Tried running videocard in second PCIe 16x slot: same issue
- Tried swapping videocard with identical 6950 from my brother: same issue
- Tried swapping powersupply with identical OZC700MSXP from brother: Same issue
- Tried running videocard off one power supply and motherboard off the other: same issue
*** Ruling out Power issue
- Tried installing old Radeon x800xl videocard: same issue
*** Ruling out Videocard issue
- Reseated CPU: same issue
- Replaced CPU with i5 2400: same issue
*** Ruling out CPU issue
So I don't know what it is, but it seems running a discreet videocard with the memory set to 1600mhz with 4 DIMMS seems to be an issue. running it at 1333mhz seems unstable as well, but doesnt always crash/BSOD - just most of the time.
Im starting to believe its a problem with the motherboard.
