BSOD on cold start

[MaxQuest]

Hello,
I would much appreciate if someone could help me with following problem:

- I receive 4E, 1A and 7E BSOD errors on cold start during windows 7 startup. (fresh install on a newly build system)

- On first try I get 4E, on 2nd boot: 1A, then 7E. Each time I selected "normal start-up" not recovery mode.
- On 4th try, windows starts normally,.. and works perfectly, i.e. it's stable without slow-downs.
- If I reboot the system again, it starts normally. But if it was shut down for more than 5 hours, then I get those BSODs again.


System info:
OS: Windows 7 Ultimate

CPU: core i3 550
GPU: Inno3D GTX460 OC
M/B: Gigabyte GA-P55-USB3.0 rev2.0
RAM: OCZ 2x2Gb 1333MHz
HDD: WD 2x500Gb Caviar Blue + 1500Gb Caviar Green
PSU: Glacial Power 550W 80+ series

Additional Info:
- My blue caviars are connected to SATA ports controlled by M/B chipset, and they are set up in XHD Raid (here I installed the OS)
- Caviar Green + DVD-Writter are connected to SATA ports controlled by onboard SATA-II chip (and are set in AHCI mode)

This is what I see using Blue Screen Viewer:

Bigger screenshot: bsod.png

And here I uploaded minidumps themselves: http://www.filefactory.com/file/b562e52/n/BSOD.zip

P.S. Immediately after 1A BSOD, I've ran memtest+. And it detected few error with my RAM.
Since I have 2 planks of memory, I've taken one out. And tested each separately. No errors were detected.
After that I placed both again, just like they were before. And memtest hasn't detected any errors. That's a bit confusing.

Also I've ran chdisk. And again no errors were detected.

 

[davcon]

Sounds like hdd issues..corrupt/missing files or bad sectors
Run sfc/scannow
http://www.sevenforums.com/tutorials/1538-sfc-scannow-command-system-file-checker.html

 

[MaxQuest]

Strange. I have run this command in the morning, and no errors were detected.
But now, during rescan, it reported that part of files are corrupted. And since part of them could not be restored, I will reinstall OS again.

 

[davcon]

Strange. I have run this command in the morning, and no errors were detected.
But now, during rescan, it reported that part of files are corrupted. And since part of them could not be restored, I will reinstall OS again.

That's your best option.
Been there done that!

 

[MaxQuest]

Damn,
Have reinstalled the system (a clean, fresh installation).
And after night I get those BSODs again. On first launch: 1A - Memory Management. On second: 4E - PFN List Corrupt.

After running memtest+ for 20min, no errors were detected.

By the way, I have placed my memory planks into 1st and 3rd slot, counting from CPU.
But in the M/B user's manual and on PCB, memory slots are counted as [2, 1],[4, 3]. And gigabyte "recomends" placing memory into slots 1 and 3 (for dual channel).
The funny thing is, that if I place the planks as recommended, the PC wont boot at all.

 

[daship]

20 min memtest = fail. run memtest86 overnight.

 

[davcon]

^+1 i also agree 20 min isn't nearly long enough
Plus you should be testing one stick at a time.

 

[MaxQuest]

I am speechless. After running for more then one hour, memtest started to catch one error after another.
Result: 327936 errors in approximately 3 hours and a half. (see screenshot)
This is what I got, when runing memtest+ on both sticks.
Now, I removed one, and started the test again.

Btw I looked in bios on the memory voltage settings:
1. DRAM Voltage: auto: 1.584V
2. Profile DDR Voltage: 1.5V (unmodifiable)
3. Profile QPI Voltage: 1.1V (unmodifiable)

While my memory (OCZ 1333Mhz Gold Series Low Voltage) requires 1.65V. Should I increase DRAM Voltage?
P.S. I can't set exactly 1.65, because of 0.2 step. But I guess going above 1.65 could degrade the CPU's memory controller?

 

[davcon]

I am speechless. After running for more then one hour, memtest started to catch one error after another.
Result: 327936 errors in approximately 3 hours and a half. (see screenshot)
This is what I got, when runing memtest+ on both sticks.
Now, I removed one, and started the test again.

Btw I looked in bios on the memory voltage settings:
1. DRAM Voltage: auto: 1.584V
2. Profile DDR Voltage: 1.5V (unmodifiable)
3. Profile QPI Voltage: 1.1V (unmodifiable)

While my memory (OCZ 1333Mhz Gold Series Low Voltage) requires 1.65V. Should I increase DRAM Voltage?
P.S. I can't set exactly 1.65, because of 0.2 step. But I guess going above 1.65 could degrade the CPU's memory controller?

Those voltage settings are mobo default which is ideal for running a memtest.
It's worth a try but i highly doubt it will make a difference.
The ram is crap! it's time for an RMA.
I knew OZC was bad ! but that's ridiculous!
OZC knows this that's why they are out of the ram business.
Too bad and good luck with your rma!

 

[MaxQuest]

Regarding voltage, indeed there was no difference.
When setting explicitly 1.64v, BIOS automatically decreases it to 1.584, because otherwise it fails to boot.
- Have run memtest+ for 6 hours (20 passes) only on first stick and had no errors.
- Right after that, I have substituted it with second stick, and memtest started to capture one error after another, starting with 1st pass.
- I've shut down the PC, and let it cool for 3 hours. Right now starting the test on 2nd stick again.

P.S. Indeed, it seems it's time for RMA, moreover I bought the memory just 9 days ago.
P.P.S. I still can't get why I can't place both sticks as recommended in M/B user's manual, i.e. in slots [1] and [3] for dual channel. I get endless short beeps on boot. Usually it is related to memory (beepcodes). But in user's manual it is clasified as power error. What the hell?
But If I place the sticks in 1st and 3rd memory socket, counting from CPU, then PC at least boots up. (through these sockets are actually [2] and [4] according to manual and inscriptions on PCB)

[cpu] | [memory [channel_0 [2][1]] | [channel_1 [4][3]]]


http://download.gigabyte.eu/FileList/Manual/mb_manual_ga-p55-usb3_v2.0_e.pdf

 

[MaxQuest]

I am clueless again. Just out of curiosity I've ran memtest again, only with 2nd memory stick plugged (which was failing during night).
Result: 7 hours - no errors found.

What conclusion should I make? :
- fail memory
- memtest+ is a fail test (through I doubt it)
- fail motherboard

What the... If a memory stick was generating errors starting with first memtest pass, why it didn't generated any errors next day for 7+ hours S.
I have also tried again to place memory sticks in slots [1] and [3]. And aleluia, the PC has booted up! (it never did before, unless I was using slots [2] and [4])
Through memtest immediatly started to catch a lot of errors...

 

[MaxQuest]

Damn. I have exchanged my fail RAM:
[2x2Gb OCZ 1333Mhz - gold edition low VT] for [2x2Gb A-DATA 1600Mhz gaming series]
.. but the problem remained.
The system is very unstable when both sticks are inserted.

- To be sure that it's not RAM problem, I've ran each stick separately for 5 hours in memtest+. They are both fine.
- I've tried to boot and start windows with only one stick plugged-in. Each time I was putting the stick in a new slot. And in all 8 combinations (4 slots, 2 sticks), windows was starting successfully and worked stable.

- But if I insert both sticks at a time, then windows doesn't start at all. Moreover trying to run memtest, freezes after couple of seconds.
I've tried:
- loading "Safe Defaults"
- changing DRAM voltage: 1.584 (auto), 1.5 (normal), and 1.648
- changing RAM frequency: 1333 (default) and 2x800 (XMP enabled)
- increasing timings
- different combinations of sockets:
--- [1]+[3] (recommended dual channel)
--- [2]+[4]
--- [1]+[2]
--- [1]+[4]
- reinstalling windows
- I've also tried downloading and installing trials of other windows 7 editions

Each time, PC boots ok, but either on windows load, either during first minute of memtest+ it freezes. And on first restart it emits continuous short beeps.
I spent a lot of time, and still have no idea what is the problem.
Is it motherboard? Memory power? Or the CPU's memory controller?

P.S. I have returned today: cpu, psu and motherboard.
I could either wait for testing/service period, either get money back. I have selected second option, because It could take up to 21 working days. But I am still curious what could be the problem?

P.P.S. Now I am looking for those components again...
Should I risk and take Gigabyte's GA-P55-USB3 again? =) Or AsRock P55 Pro/USB3 might prove to be more stable?