CMT16GX3M4X2133C9 + Alienware Aurora-R3 - Unhappy Marriage

I recently purchased 4 sticks of CMT16GX3M4X2133C9 Corsair memory to replace the 8GB of lower-quality Micron memory in my new Alienware Aurora-R3. The system is running an Intel i7-2600k processor @ 4.1GHz on a P67 chipset. I'm using the standard XMP timings of 9-11-10-27-2T.

 

Unfortunately, I am contending with a number of BSOD in an otherwise stable install of Windows 7 since making this upgrade. The memory successfully completed two passes in Memtest+ 4.20 at the above timings.

 

I am wondering if this memory might require different timings (or other BIOS tweaks) to make a happy bedfellow of my Aurora-R3?

 

Thank you in advance for any help you might offer.

 

Best,

 

Jason

 

EDIT: As an added data point, there does seem to be a correlation with voltage and stability at these timings. At 1.5v, the system is very unstable. At 1.6v, the system appears relatively stable but apps occasionally throw errors; OCCT runs < 10 min before pitching a fit. At 1.65v OCT runs for 20-30 min before choking. I am loathe to take the voltage any higher than this until I know what the over-voltage limits are on these sticks.

 

EDIT: Also tried XMP profile timings of 9-12-11-28-2T@1.65v and system would not POST.

I think I'll recant just a bit. There doesn't really seem to be any real correlation between stability and voltage in this configuration. I think I just had a "lucky" run. Back to running crappy Micron China-RAM for now. For some reason I got it into my head that if China-RAM runs like a champ with the Aurora-R3's factory config then installing faster beefier Corsair Vengeance memory with twice the perf. specs should "work" just swimmingly out of the box. :\ Obviously, the new Sandy Bridge universe is more complicated than this.

Well, first , 2133mhz is an awfull huge memory overclock for your CPU. From what you have posted it looks like your CPU just wont overclock the memory that high. Have you tried lowering the frequency to something like1866mhz?

 

Or have you tried adding memory controller voltage ? VCCIO?

I tried raising both the controller and DRAM voltage - thinking maybe I was in violation of the sacrosanct 0.5v rule - without success. Unfortunately, the BIOS on this system doesn't give the kind of insight into actual voltages that my older Maximums Extreme did. This BIOS just gives the option to "overvolt" the controller anywhere from 0 to 0.6v in 40mv increments. I was loathe to try anything over +300mv, since anything over 1.4v makes me nervous.

 

China-RAM is running 9-9-9-25 @ 1333. Maybe, it's as you say, the IMC just can't keep up and/or I'm running up against a limitation of the P67. I'll try an underclock and see how I fare. Maybe I purchased RAM well beyond the means of my system which would be somewhat regrettable.

 

J

 

PS. Is there a better app other than HWMonitor for monitoring MB voltages that people are using? I can't seem to find one that correlates to the IMC/VTT/VCCIO voltage. VIN3 is the closest I could find (shows 1.31v - always) and doesn't seem to go up or down in response to changes in the BIOS. I have fears of smoking the IMC going in blind as I am (I'm assuming a base of 1.1v but would rather have a real measurement).

 

UPDATE: Well, it looks like you might be on to something. This memory may be just "a bit too big for its britches". I took your advice and throttled back to 1866MHz and the system seems to be in a much happier place (at stock voltages and XMP timings). Thanks for the tip. I'd still like to know what VIN3 corresponds to in HWMonitor on this board. It seems to vary depending on the active memory configuration. It tends to start high at boot (around 1.6v or so) and gradually floats down to 1.31v where it settles.

PS. Is there a better app other than HWMonitor for monitoring MB voltages that people are using? I can't seem to find one that correlates to the IMC/VTT/VCCIO voltage. VIN3 is the closest I could find (shows 1.31v - always) and doesn't seem to go up or down in response to changes in the BIOS. I have fears of smoking the IMC going in blind as I am (I'm assuming a base of 1.1v but would rather have a real measurement).

Not really, You are not going to get true measurements with a software based app anyway. The ONLY way to get true measurements is with a multimeter. Bios would be the next best thing above the meter.

 

Well, it looks like you might be on to something. This memory may be just "a bit too big for its britches". I took your advice and throttled back to 1866MHz and the system seems to be in a much happier place (at stock voltages and XMP timings). Thanks for the tip. I'd still like to know what VIN3 corresponds to in HWMonitor on this board. It seems to vary depending on the active memory configuration. It tends to start high at boot (around 1.6v or so) and gradually floats down to 1.31v where it settles.

Yeah, it happens alot. People see that a MB may support high memory speeds and not realize that the CPU plays a huge role as to whether or not you can actually run those speeds. But let it go at 1866 for a while and see how you fare!

 

As far as the voltage goes, i'm not entirely sure about that. So i cant give you an answer to that one. Maybe RamGuy would know, or Synthohol. I just dont want to give you an answer that may not be right.