F3-10666CL8D-4BRM Cold Boot BSOD.

[Manacla_AR]

Hi!, thanks for letting me write in this forum. I apologize for any grammar or spelling mistake, mind you im 15 and from Argentina.

First of all, I would like to inform you about my specs and what has been happening with my PC=
CM690
H55M-USB3 (The Bios is already updated to the last version)
I5 750 2.6ghz STOCK
HD5850 XFX XX 725/1125 OC'd from factory.
2x2 GSKILL 2gb f3-10666cl8d-4gbrm CL8-8-8-24 ddr-1333 1.5v (this is what the blister says)
WD Black 500gb 7200rpm.

Ok, let me tell you a little thing first. The GSKILL memories werent the ones I always had. Before getting this pair of GSKILL 2x2gb ram, I got a 2x2gb OCZ, but the 1st problem we (my father and I) had with them is that a stick came faulty from factory, and then the RMA replacement was....mmm...."uncompatible" with my MOBO.
So, we went to our tech dealer (I live in the countryside of Capital Federal, which is 30km away) in Capital Federal (http://tecnohards.com.ar/) and asked them to reeplace the RAM. We got the already mentioned RAM instead of another pair of OCZ.
Ok, so, the settings in the bios are:
Timings= 8-8-8-24
Voltage= Auto (It's set to 1.5v).
Frequency= 1333mhz.
Here is a screenshot from CPU-Z so you can see everything in detail.


Ok, so now that I've briefed you about my system and specs I believe its time to move onto the problem:
PLEASE NOTICE THAT THIS STARTED TO HAPPEN WITH THE GSKILL MEMORIES, THE OCZ MEMORIES DIDNT BSOD ON COLD BOOT.

The problem is that whenever I turn on the computer after leaving it for X>5 hours, it BSODS with a 0x0000001a MEMORY_MANAGEMENT BSOD.

It's a damn axiom, it WILL happen all times, sooner or later.

Now, the odd thing is that, after several restarts (sometimes it takes one restart, sometimes it takes X>1 restarts) the PC WONT BSOD AGAIN.


So, you might say, "Dude, have you checked if its the MOBO, the PSU, the GPU, the HD?"
Well, the answer is: "Hell yeah, I've checked and double checked everything".

I've ran several OCCT PSU/CPU tests, all passed, no problem.
I've ran several Prime95 tests, all passed, no problem.
I've ran several OCCT GPU tests and played BF3 for more than 6 hours, no problem.
I've ran several times the WD Diagnose Tool in extended mode, all passed, no problem.
I've even also ran several times the MEMTEST for more than 8 hours, all pssed, no problem.

Whoa, you just said you ran OCCT CPU, Prime95 and MEMTEST several times and they didnt show any error, so, why do you put the blame on the RAM?....

Well....=
http://forums.anandtech.com/showthread.php?t=2039083
http://www.gskill.us/forum/showthrea...ight=cold+boot

The guys were having the SAME EXACT problem as me.
Plus, if I think about it:
-This didnt happen with the OCZ sticks.
-If the BSODs are due to a faulty PSU/GPU/CPU, then they would have to happen at all times, not just on cold boot.
-This started just when I got the GSKILL sticks.

What's the posible solution for this? Do I have to get another brand different than GSKILL? Do I have to RMA them? Do I have to up the voltage in order for them to get "up" more faster?.

Hope you can help me. If I wasnt clear in something, I urge you to point it out ASAP. This is a priority for me and my father.

EDIT:
I forgot to add something. I found out this while searching for memory info the first day I got them=

Gigabyte GA-P55 UD4
Gigabyte GA-P55 UD4P
Gigabyte GA-P55 UD5
Gigabyte GA-P55 UD6
Gigabyte GA-H55M-UD2H

ASRock M3A790GXH/128M

The H55M-USB3 is not there....

[Tradesman]

I'd try raising the QPI/VTT voltage just a bit ( + .02/.03) and/or possibly the DRAM (+.04), they may not be getting the true voltage, saw something a while back on this mobo about that

[Manacla_AR]

How do I do that?
I've never seen the QPI/VTT voltages and DRAM voltages values,

I'd try raising the QPI/VTT voltage just a bit ( + .02/.03) and/or possibly the DRAM (+.04), they may not be getting the true voltage, saw something a while back on this mobo about that

Also, Have you read the whole post? Do you know what happens after several restarts? It makes no sense at all. If they arent getting their true voltage, then it should BSOD at all times.

[Tradesman]

If the mobo isn't reading the sticks right through the BIOS it may be sending default values to them, which may not be true to the sticks. The voltages can be found in the BIOS settings. Might also want to change CR Command Rate to 2T, the sticks are designed for 2T as is shown in the XMP settings under the SPD tab of your CPU-Z image, yet in the memory tab (which shows what they are actually running at) shows the mobo/BIOS has it at 1T. The latest BIOS, F11 is over a year old so it appears they have quit updating it with newer sticks and settings. Link to BIOS updates:

http://www.gigabyte.com/products/pro...?pid=3444#bios

[Manacla_AR]

Have you seen the JEDEC #3? CHECK OUT THE TIMINGS.
They're wrong, how come JEDEC#3 puts them as 9 9 9 24 when its 8 8 8 24?

Something really wrong here, and I doubt its the mobo....
Im going to check the settings in the BIOS.

EDIT: I manually set the voltage to 1.560v, is that right? I mean, I can only choose 1.5v or 1.56v
EDIT2:FOUND THE QPI/VTT SETTINGS, IT'S SET AT 1.100V, NEXT OPTION IS 1.150V.
DO I KEEP IT THERE OR I CHANGE TO 1.150V?
I'VE MANAGED TO PUT THE CR TO 2T.
LET'S SEE HOW IT COMES OUT.

[Tradesman]

You need to know what CPU-Z is doing - what it's doing in the SPD tab is reading specific settings off the SPD programmed into the stick (their labeling sucks) what it is saying in this case is that at TIMINGS of
8-8-8-22 it will run at 592 (or 1184) which is the closest it's programmed for to a JEDEC standard, in this case 8-8-8-24.....another oddball on the SPD tab is where they make an identification of "MAX BANDWIDTH) which has nothing to do with the data it pulls, the data listed there is the base default for boot the sticks use, which generally these days is 1333 or 1600 for DDR3...contrary to many who claim that is the true speed of the chips, which has absolutely nothing to do with it. All sticks need a setting for a basic default boot, since the mobos go to a basic default when they see sticks for the first time

[Manacla_AR]

You need to know what CPU-Z is doing - what it's doing in the SPD tab is reading specific settings off the SPD programmed into the stick (their labeling sucks) what it is saying in this case is that at TIMINGS of
8-8-8-22 it will run at 592 (or 1184) which is the closest it's programmed for to a JEDEC standard, in this case 8-8-8-24.....another oddball on the SPD tab is where they make an identification of "MAX BANDWIDTH) which has nothing to do with the data it pulls, the data listed there is the base default for boot the sticks use, which generally these days is 1333 or 1600 for DDR3...contrary to many who claim that is the true speed of the chips, which has absolutely nothing to do with it. All sticks need a setting for a basic default boot, since the mobos go to a basic default when they see sticks for the first time

Ahhh, good good.Thanks

Have you read the EDITS?

[Tradesman]

Sorry, try the 1.56 on DRAM first then try adding to the QPI/VTT to 1.15 if needed

[Manacla_AR]

Sorry, try the 1.56 on DRAM first then try adding to the QPI/VTT to 1.15 if needed

Yes, upping the voltage has managed to "erase" the BSODS. But the things that happened before them, such as programs crashing, explorer.exe crashing and so on, still happen. And they're still fixed after restarting the machine.
This might be a corruption caused by the memories when they were at 1.5v.

[Manacla_AR]

Ok, there are no more BSODs, but now the machine stutters like never before, I've got freezes that last up to 20-25 seconds!.

[Tradesman]

Could be corruption of the OS from the BSODs and the system not being able to shut files before power off...could try running CHKDSK /R from the command prompt and the SFC /Scannow from the Command prompt

[Manacla_AR]

Tradesman, seems like upping the voltage didnt work at all, right now I did a cold boot (12:33 pm, turned it off yesterday at 10pm) and it BSOD'ed right after Firefox (I uninstalled Chrome just to see if it was file corruption) crashed."Mem_manag"... again.
So, like you said, I upped the QPI/VTT voltage to 1.150v.
What Im doing next it to reinstall Windows and see how it goes. Let's hope it doesnt BSOD on cold boot again.

[Tradesman]

Let us know what happens

[Manacla_AR]

Let us know what happens

Im searching for the W.32 Bits DVD.
Meanwhile, Im curious, what's up with the QPI/VTT voltage and the RAM?
Thanks for helping me btw.

[Tradesman]

As far as how they relate to each other? The QPI/VTT supplies the voltage to the MC (Memory Controller) and at times it doesn't provide enough to handle the number of sticks or the amount of DRAM..it's sort of a BIOS thing...if you look at a page of BIOS updates for a mobo, they'll normally list 1 or 2 main ideas addressed in the update, but there are generally a ton of DRAM updates also....It's one of the big reasons you often see folks that update the DRAM on what I think of as an 'orphaned' mobo (one they are no longer supporting with BIOS updates), and the people use a new brand/model of sticks, they run into the BIOS not sure of how to approach the setup