Solved Fresh Build - Constant BSOD - Multiple Codes - 3b, 50, a

[LP81]

Hi All,

After my last rig decided to stop working I decided to build a new one and start fresh. Would randomly turn off (not BSOD) and take multiple tries to turn back on

My new rig is based off a PCGamer build, so I know components should work well together:

OS: Windows 7 Ultimate x64
CPU: Intel Core i5-2500K @ 3.30Ghz
Motherboard: Gigabyte GA-Z68AP-D3 Intel Z68
Memory: 2x Corsair 4GB Dual Channel Corsair DDR3 for Intel Core i5
Graphics Card(s): GeForce 8800 GT (waiting on Sapphire Video Card Radeon HD)
Sound Card: Creative Sound Blaster X-Fi XtremeGamer Fatal1ty Pro 7.1
Monitor(s) Displays: Gateway 21" Monitor and output to TVHard Drives: Western Digital WD1002FAEX 1TB Caviar Black - 7200 RPM, 64MB, SATA 6Gbs
PSU: Ultra X4 750 Watt Modular PSUCase: COOLER MASTER COSMOS 1000 RC-1000-KSN1-GP
Cooling: 4- 120MM Case Farns plus stock Intel CPU Cooler

This issue seems to be isolated to the Win 7 itself.

I kept my previous Hard Drives and was able to boot up XP and it was 100% stable. Loaded up my Win7-64 install and nothing my BSOD dumps. Hardware was no different.

So I decided to buy a new hard drive, and start completely from scratch with a fresh drive, fresh Win7 Install and everything. The only 4 parts above that are "old" are the Original Case, Monitor, Graphics Card (as the one I ordered hasn't shown up yet) and the SoundBlaster Audio Card

The rest of the items are 100% new, fresh out of the box.

I have run the Win 7 Memory Diagnostic Test and it gave a clean bill of health.

I installed Win7 this morning and in the last 2 hours I have had 5 dumps. Any help is much appreciated. I am working on uploading the zip, but since it crashes so often, it is being troublesome to pull down the exe and run it. Submitting this thread from laptop while I try to get the info from the main computer.

 

[LP81]

zip attached

Zip is attached

 

[Vir Gnarus]

First rule: Windows 7 Memory Diagnostic Test is not recommended. Memtest86+ is the de facto standard for memory testing. Run it 7 passes minimum on your memory.

When attempting to gather information, have you tried going into Safe Mode with Networking to do it? You may have better luck there.

 

[LP81]

Dumps seems to be consistent regardless of Normal Mode vs Safe Mode.

I'll attempt to download the Memtest86+ listed above, but as I am not able to keep the system up very long, I don't know if I'l be able to pull the ISO down, get it mounted/burned before it crashes again.

 

[Vir Gnarus]

Ah, ok, I received your JCGriff Report and looking through it now. So far, nothing exactly definitive, but I can see on the crashdumps that your CPU registers are being messed up. The highest octet for them is being manipulated in which the same 2 bytes are being changed. This is occurring when a memory address is residing in the CPU register, which is why when the system tries to access memory based on the cpu register by the now defunct address, the system detects the issue and BSODs.

I'm not able to determine what is corrupting these two bytes, whether it's a driver or if it's a piece of hardware. I can never debug this stuff with minidumps because every time I run into a brick wall because the minidump doesn't hold the information I need. Only a kernel dump (MEMORY.DMP located in your Windows directory) or larger is enough to work with.

However, this may very well be hardware. The reason why you may not be getting the crashes in Windows XP, is because Windows XP is 32-bit while your Windows 7 is 64-bit. The CPU registers used by 32-bit and 64-bit OSes are the same, but the 32-bit OSes use only the first half of the registers while the 64-bit OSes uses them fully. Since this corruption I'm seeing only seems to occur in the highest octet (half) of the CPU registers, it's no wonder why Windows XP would not crash, because Windows XP wouldn't access the highest octet for those registers even if it tried (which it won't, because it can't see anywhere beyond 32 bits).

Given that this is a brand new system, I would not doubt there's something funky going on with your hardware as there's always the occasion hardware can be received in a defective state.

If you wish, you may do some more hardware tests. I prefer running Memtest86+ before as mentioned. In addition, in Safe Mode for Windows 7, try running Prime95 overnight. Run it on Torture Test, running it on Small FFTs for one night, Large FFTs for another night, and finally Blend for a 3rd night. Each test does things differently and if there's any crashing that occurs on any of them make sure to record such. Depending on which tests crash we can better ascertain the hardware involved that did it, be it RAM or CPU.

Lastly, a couple of other things. If you have anything on your system overclocked, please reset it all to factory defaults. That should be a given for any system that you find is starting to experience instability. In addition to this, confirm if your CPU is not overheating. You can do so in the BIOS or with some temp software like HWInfo32. Make sure to check Sensors only at startup if you plan on using it. Also be aware that motherboard software is commonly buggy and I've found is notorious for causing errors that appear hardware-based. Uninstall any and all software that came with your motherboard. Only install the latest drivers (and BIOS) for the components to your motherboard (LAN, chipset drivers, drive controller, etc.). If the vendor suggest installing drivers for USB or any other type of port interface, do not install them. They are just ugly software posing as drivers.

If you wish for my personal take on this, I say that the CPU is to blame here. The errors seem too consistent to be from a PSU or motherboard, as well as RAM. However, it's best to look further into this before going caveman and start throwing stuff out and swapping them with new parts.



Analysts:

Ugly registers below, each printout for each crashdump where it's discovered:

TRAP_FRAME:  fffff88008986d00 -- (.trap 0xfffff88008986d00)
NOTE: The trap frame does not contain all registers.
Some register values may be zeroed or incorrect.
rax=fffffa80090a6de0 rbx=0000000000000000 rcx=ffff[COLOR=Red]95[/COLOR]80071754e0
rdx=0000000000000040 rsi=0000000000000000 rdi=0000000000000000
rip=fffff80002af6142 rsp=fffff88008986e90 rbp=fffff80002c46e80
 r8=0000000000000000  r9=fffff88008986f90 r10=0000000000000000
r11=0000000000000000 r12=0000000000000000 r13=0000000000000000
r14=0000000000000000 r15=0000000000000000
iopl=0         nv up ei pl zr na po nc
nt!MiAddViewsForSection+0x122:
fffff800`02af6142 48894108        mov     qword ptr [[COLOR=Red]rcx+8[/COLOR]],rax ds:ffff9580`071754e8=????????????????

__________________

TRAP_FRAME:  fffff88007bf5170 -- (.trap 0xfffff88007bf5170)
NOTE: The trap frame does not contain all registers.
Some register values may be zeroed or incorrect.
rax=ffff[COLOR=Red]92[/COLOR]a002b0f178 rbx=0000000000000000 rcx=fffff8a002b0f178
rdx=0000000000000011 rsi=0000000000000000 rdi=0000000000000000
rip=fffff80002a63d84 rsp=fffff88007bf5300 rbp=fffffa8007dfe010
 r8=0000000000000000  r9=fffff88007bf53d0 r10=0000000000000000
r11=fffff88007bf54c8 r12=0000000000000000 r13=0000000000000000
r14=0000000000000000 r15=0000000000000000
iopl=0         nv up ei ng nz na po cy
nt!FsRtlLookupPerStreamContextInternal+0x7c:
fffff800`02a63d84 48396810        cmp     qword ptr [[COLOR=Red]rax+10h[/COLOR]],rbp ds:ffff92a0`02b0f188=????????????????

__________________


CONTEXT:  fffff8800553ebe0 -- (.cxr 0xfffff8800553ebe0)
rax=ffff[COLOR=Red]00[/COLOR]8009471ea8 rbx=0000000000032403 rcx=0000000000000010
rdx=fffff8a00020ad20 rsi=fffffa80073ad080 rdi=0000000000000400
rip=fffff80002b8e78c rsp=fffff8800553f5c0 rbp=0000000000000000
 r8=fffffa80094a0003  r9=fffffa80073ad048 r10=fffffa8009471df0
r11=fffffa800947bdd0 r12=0000000000000000 r13=fffffa80073ad000
r14=fffff8a00020acf0 r15=0000000000000002
iopl=0         nv up ei ng nz na po nc
cs=0010  ss=0018  ds=002b  es=002b  fs=0053  gs=002b             efl=00010286
nt!MiEmptyPageAccessLog+0xdc:
fffff800`02b8e78c 488b08          mov     rcx,qword ptr [[COLOR=Red]rax[/COLOR]] ds:002b:ffff0080`09471ea8=????????????????

__________________

CONTEXT:  fffff88008babd50 -- (.cxr 0xfffff88008babd50)
rax=0000000000539ea0 rbx=fffffa8007373860 rcx=0000000000000000
rdx=0000000000000000 rsi=fffffa80075fa460 rdi=ffff[COLOR=Red]04[/COLOR]8006fa0a10
rip=fffff80002ab1b46 rsp=fffff88008bac730 rbp=0000000000000001
 r8=fffffa80073ea148  r9=0000000000000000 r10=0000000000000000
r11=0000000000000000 r12=0000000000000000 r13=fffff8a00c1f3a01
r14=fffff8a00c1f3a00 r15=0000000000000000
iopl=0         nv up ei pl zr na po nc
cs=0010  ss=0018  ds=002b  es=002b  fs=0053  gs=002b             efl=00010246
nt!AlpcpQueueIoCompletionPort+0xb6:
fffff800`02ab1b46 44897f28        mov     dword ptr [[COLOR=Red]rdi+28h[/COLOR]],r15d ds:002b:ffff0480`06fa0a38=????????

__________________

CONTEXT:  fffff88003184d30 -- (.cxr 0xfffff88003184d30)
rax=0000000000000002 rbx=[COLOR=Red]00110a80051e6b20 [/COLOR]rcx=fffff88003185670
rdx=fffffa8000000000 rsi=0000000000000000 rdi=0000000000000048
rip=fffff80002878285 rsp=fffff88003185710 rbp=fffffa80073e91c0
 r8=0000058000000000  r9=0000000000000000 r10=0000000000000000
r11=fffff880031856c8 r12=0000000000000000 r13=0000000000000001
r14=fffffa80073e94e8 r15=fffff8a007db24b8
iopl=0         nv up ei pl nz na pe nc
cs=0010  ss=0018  ds=002b  es=002b  fs=0053  gs=002b             efl=00010202
nt!MiFlushSectionInternal+0xb1c:
fffff800`02878285 f00fba6b1000    lock bts dword ptr [[COLOR=Red]rbx+10h[/COLOR]],0 ds:002b:00110a80`051e6b30=????????

Last crashdump has a register that's just a mess, but the rest have just that single word (2 bytes) that's defective. Regardless of which cpu register it is, that thing always seems to pop up in the same place in the register. This may be software based or a problem with the CPU, or even RAM. Again, I cannot tell without a kernel dump or more.

 

[LP81]

Thanks,

I am running Memtest86+ right now - it's completed one pass so far no errors, but I will let it run for a minimum of 7 passes as you recommended (screenshot attached for reference)

No overclocking has been done, straight up build, loaded Win 7 and BSOD right away.

Only Drivers I installed from the CD were LAN, chipset, SATA & USB. I skipped their HD Audio Drivers, Utilities & Anti Virus. Once the Memtest86+ is completed, I'll try to remove the USB "drivers" as suggested.

I have multiple temp gauges throughout the case, Average temp is about 80•F right now

I did have to RMA the Hard Drive sent to me, I hope not to RMA the processor as well.

 

[Vir Gnarus]

Yeah, there's definite hardware issues here. I personally am pointing blame at the CPU, but we're just gonna make sure here. These hardware tests may not show up anything at all, but that doesn't mean the hardware you tested isn't to blame, it just means the tests weren't thorough enough to discover it. Prime95 may be able to stress test a CPU, but it only uses a subset of instructions that the CPU is capable of, as well as not all the caches and registers available to it.

Sometimes it may have to come to swapping hardware when building a new system. The hardware tests are there to help ascertain which hardware piece is buggering out, but they're not fail proof. Also, PSU and Motherboard issues are notorious in that not only can they either show up problems in multiple hardware tests, they can also show up in no hardware tests, causing all those tests to pass and leave the person scratching their head wondering what exactly is causing the issue. There are no tests for the PSU or Motherboard other than buying a multimeter and a specialized motherboard diagnostic kit.

 

[James7679]

May I ask what frequency you're actually running the RAM at? Do you have XMP enabled in the BIOS? 0x3b, 0x0a and 0x50 play some common factors here. Voltage instability between the CPU and Memory.

Please, if you don't mind, upload the first tab and fifth tab from CPUz.

 

[LP81]

I am running the RAM at whatever the default setting is.

The Memtest86+ is running right now and is 73% thru the 4th pass so I am in a holding pattern til it finishes.

Per the memtest86+ screen it says "Settings : RAM: 645 Mhz (DDR3-1291) / CAS : 9-9-9-24 / Dual Channel"

Also below that with the "Memory SPD Informations" it has a line for each stick of memory:

- Slot 0 : 2048 MB PC3-10600 - Corsair CMX4GX3M2A1600C9 *XMP*
same info for slot 1, 2 3 as well

Once i have finished the 7 recommended passes I can go into the BIOs to look up more info

 

[James7679]

I am running the RAM at whatever the default setting is.

The Memtest86+ is running right now and is 73% thru the 4th pass so I am in a holding pattern til it finishes.

Per the memtest86+ screen it says "Settings : RAM: 645 Mhz (DDR3-1291) / CAS : 9-9-9-24 / Dual Channel"

Also below that with the "Memory SPD Informations" it has a line for each stick of memory:

- Slot 0 : 2048 MB PC3-10600 - Corsair CMX4GX3M2A1600C9 *XMP*
same info for slot 1, 2 3 as well

Once i have finished the 7 recommended passes I can go into the BIOs to look up more info

Sounds great. FYI, this RAM is meant to run with XMP enabled. Yes, it does matter. Underclocking RAM usually causes problems. Here is CPUz, if you don't have it.

 

[LP81]

Up to 5 successful passes on Memtest86+

here's a small snippet from the Motherboards manual: Gigabyte GA-Z68AP-D3 | Manual - Page 37

Extreme Memory Profile (X.M.P.), System Memory Multiplier (SPD), Memory Frequency(Mhz)
The settings under the three items above are synchronous to those under the same items on the Advanced Frequency Settings menu.

Performance Enhance
Allows the system to operate at three different performance levels.
Standard -- Lets the system operate at its basic performance level.
Turbo -- Lets the system operate at its good performance level. (Default)
Extreme -- Lets the system operate at its best performance level.

DRAM Timing Selectable (SPD)
Quick and Expert allows the Channel Interleaving, Rank Interleaving, Channel A Timing Settings, and Channel B Timing Settings items to be configurable. Options are: Auto (default), Quick, Expert.

Profile DDR Voltage
When using a non-XMP memory module or Extreme Memory Profile (X.M.P.) is set to Disabled, this item will display as 1.5V. When Extreme Memory Profile (X.M.P.) is set to Profle1 or Profle2, this item will display the value based on the SPD data on the XMP memory.

So it sounds like the XMP is on by default but I will verify setting after the 7th RAM testing pass

 

[LP81]

7 Passes, no errors

Went into the BIOS and checked out the Memory Frequency Tab:
X.M.P. is currently disabled - screenshot below

Also I downloaded and ran CPUz. Grabbed multiple screenshots. Computer BSODed in the middle of me saving them, this time was a d1 Error

Tab1:

Tab4:

Tab5: (all 4 sticks in the sub menu had the same info)

 

[James7679]

Try enabling XMP. Also, are you able to run Prime95 without crashing? Or Intel Burn Test?

Lastly, don't suppose you have any other hardware laying around that you can use for testing? Another 1155, or MoBo, or PSU?

 

[LP81]

Downloaded Prime95, ran the Small FFTs test and I got a BSOD after about 20 min - 3b error

PSU I know isn't the cause. I originally had a OCZ ZS Series 550W 80PLUS Bronze High Performance Power Supply on it with the same BSOD issues.

I am not 100% what caused my previous Rig to die so trying to swap in its old guts might now be advisable
Mobo - EVGA 122-CK-NF67-T1 LGA 775 NVIDIA nForce 680i
CPU - Intel Core 2 Quad Q6600 Kentsfield 2.4GHz

I will try enabling the XMP as a next step and see what happens

 

[LP81]

Turned on XMP in BIOS, it defaulted to "Turbo"
Save & Exit
The computer shut itself down, rebooted and said "The system has experienced boot failures because of overclocking or changes of voltages. Last settings will be used"

Went back in, turned XMP on again but changed it to "standard"
Save & Exit
Windows started to boot, then reset itself, and ended up on a Windows Error Recovery screen
Chose "Launch Startup Repair" and got a BSOD 7E error

Disabled XMP in the BIOS and after a lot of hullabaloo, finally got windows to start up again.

I think the problem with this step is I sent the shutdown signal, and it started to do updates then died in the middle of them causing the previous error.

I will try to re-enable XMP in "Standard" mode again and see what happens.

 

[LP81]

So with XMP turned on, and set as "standard", I can't even get to the Win7 login screen before a BSOD. This time it's a 1000007e ndis.sys BSOD dump.

I'm about to run the jcgriff again so you can get an updated list of the 14 dumps on file.

Attachment to follow

 

[James7679]

All of these codes you are getting, when placed together definitely indicate that you have voltage issues. Now, since the memory controller is no longer on the north bridge, and you are having cpu errors, as Vir Gnarus kindly pointed out, my best guess is that you have a bad processor. This is not common, but I have seen it a few times lately. It could be your MoBo, and less likely, your memory.

There is one more thing you can try. Looking at your CPUz, your idle voltage looks pretty low, this can cause instability. It's possible that you could bump your VCore offset up and it might stabilize. Try bumping the offset by .020(can't remember if that is the differential on the offset menu), or whatever the next bump up is. Reboot, and if it seems a little better, but still crashes, bump it up a little more. Oh, and leave XMP enabled when doing so. Your settings should look like this;

Note that the Performance enhance shows "Turbo" and the XMP profile should be marked "Profile 1".


Post back results.

 

[LP81]

I can edit the VCore frequency in increments of 0.005 on this mobo.

The BIOS originally read 1.240 at the time so I upped it to 1.260 as well as changed the XMP back on to Turbo

Save and Exit - Immediate dump while loading Windows (too fast to see actual dump info)

Back into BIOS, now the CPU Vcore says 1.320V as the "Normal" so I tried 1.340V

Save and Exit - Immediate dump while loading Windows (too fast to see actual dump info)

Tried going back to "Standard" on XMP with 1.64 - same thing - quick dump


On a side note, looking at the "Advanced Memory Settings" I do see that is says the System Memory Multiplier says 1333Mhz, however the RAM i have is Corsair PC3-12800 1600MHz 240-pin DDR3 Timing: 9-9-9-24 1.65v

Could that be the issue at all? (just throwing out Ideas)

 

[James7679]

No, that's fine.

Well, I will see if anyone else has further suggestions, but my best guess is to second the opinion of Vir Gnarus and advise an RMA of your CPU.

 

[Dwarf]

We've ascertained by what you have posted that there is nothing wrong with your hardware as far as a 32-bit OS is concerned (XP running without problems). We also know that you are having problems with w7 64-bit, but can you try the 32-bit version instead? W7 contains far more drivers and is compatible straight out of the box than earlier OSes (including XP), so I don't think that it is a driver issue.

If a clean install of 32-bit works without problems, but the 64-bit version causes random BSODs, then the cause of the problem must be something to do with the architecture. We can rule out most of the hardware here, as it functions identically under both bit versions. The only component that can have a bearing on this is the CPU. It is perfectly feasible to have a CPU that functions OK with 32-bit but has problems when trying to use 64-bit.

Note that when carrying out testing such as this, you should set your BIOS to its defaults, and should stop over/under clocking. You should also ensure that you deselect the option to automatically activate when online.