Solved 3b win32k BSOD from hardware upgrades

[Tezzick]

G.Skill is high quality for Intel and AMD and are interchangeable so long as RAM specs for the board are met. Your current RAM should be reset to the timings shown in this snip, from your system.

View attachment 244585

Check you manual on how to access the RAM settings, likely under an advanced sub-menu. Note that the 2T sometimes is called 2N, it is the same thing.

Thanks, good to know. I thought I was fine buying the G.Skill, especially with the reviews.

I'll get that switched over right now and see how that runs.
Does that crash file point to memory, and this could be a fix, or just a good thing to change?

Thanks for your help, Briton, by the way. Going at this for over a week and its great to finally be heading in the right direction, haha.
You too koolkat!

 

[Britton30]

I don't read the dumps. The RAM settings probably won't get rid of BSODs but the system will run a little better.

Kool Kat will check on the dumps for you.

 

[koolkat77]

Can you keep your graphics card out until we get your RAM sorted. This dump was a random crash but showed errors with your graphics card drivers later.

STOP 0x0000000A: IRQL_NOT_LESS_OR_EQUAL    

Usual causes:  Kernel mode driver, System Service, BIOS, Windows, Virus scanner, Backup tool, compatibility
BugCheck A, {fffffa7fffffffd0, 2, 1, fffff80002a8e87c}

Probably caused by : memory_corruption ( nt!MiInsertPageInFreeOrZeroedList+54c )

CUSTOMER_CRASH_COUNT:  1

DEFAULT_BUCKET_ID:  WIN7_DRIVER_FAULT

BUGCHECK_STR:  0xA

PROCESS_NAME:  opera.exe

STACK_COMMAND:  kb

FOLLOWUP_IP: 
nt!MiInsertPageInFreeOrZeroedList+54c
fffff800`02a8e87c 498930          mov     qword ptr [r8],rsi

SYMBOL_STACK_INDEX:  3

SYMBOL_NAME:  nt!MiInsertPageInFreeOrZeroedList+54c

FOLLOWUP_NAME:  MachineOwner

MODULE_NAME: nt

DEBUG_FLR_IMAGE_TIMESTAMP:  503f82be

IMAGE_NAME:  memory_corruption

FAILURE_BUCKET_ID:  X64_0xA_nt!MiInsertPageInFreeOrZeroedList+54c

BUCKET_ID:  X64_0xA_nt!MiInsertPageInFreeOrZeroedList+54c

Followup: MachineOwner
---------
fffff880`09c12608  fffff880`04088217*** WARNING: Unable to verify timestamp for atikmpag.sys
*** ERROR: Module load completed but symbols could not be loaded for atikmpag.sys
 atikmpag+0x12217

Insert your graphics card in another PC and see if anything weird happens. You can also take the furmark and video memory test. Make sure nothing is wrong with the "new" graphics card

 

[Tezzick]

Can you keep your graphics card out until we get your RAM sorted. This dump was a random crash but showed errors with your graphics card drivers later.

STOP 0x0000000A: IRQL_NOT_LESS_OR_EQUAL    

Usual causes:  Kernel mode driver, System Service, BIOS, Windows, Virus scanner, Backup tool, compatibility
BugCheck A, {fffffa7fffffffd0, 2, 1, fffff80002a8e87c}

Probably caused by : memory_corruption ( nt!MiInsertPageInFreeOrZeroedList+54c )

CUSTOMER_CRASH_COUNT:  1

DEFAULT_BUCKET_ID:  WIN7_DRIVER_FAULT

BUGCHECK_STR:  0xA

PROCESS_NAME:  opera.exe

STACK_COMMAND:  kb

FOLLOWUP_IP: 
nt!MiInsertPageInFreeOrZeroedList+54c
fffff800`02a8e87c 498930          mov     qword ptr [r8],rsi

SYMBOL_STACK_INDEX:  3

SYMBOL_NAME:  nt!MiInsertPageInFreeOrZeroedList+54c

FOLLOWUP_NAME:  MachineOwner

MODULE_NAME: nt

DEBUG_FLR_IMAGE_TIMESTAMP:  503f82be

IMAGE_NAME:  memory_corruption

FAILURE_BUCKET_ID:  X64_0xA_nt!MiInsertPageInFreeOrZeroedList+54c

BUCKET_ID:  X64_0xA_nt!MiInsertPageInFreeOrZeroedList+54c

Followup: MachineOwner
---------
fffff880`09c12608  fffff880`04088217*** WARNING: Unable to verify timestamp for atikmpag.sys
*** ERROR: Module load completed but symbols could not be loaded for atikmpag.sys
 atikmpag+0x12217

Insert your graphics card in another PC and see if anything weird happens. You can also take the furmark and video memory test. Make sure nothing is wrong with the "new" graphics card

Well I just tired doing what Britton suggested with the different timings.
Changed it to 9-9-9-25, crashed, so I re-read the ram packaging and it says 9-9-9-24, so I changed it to 24 and changed the row cycle time to 33... not sure if I should have changed it to 33 but thats the snip it that was highlighted int he picture.

I then worked fine on the computer for a good few hours, until another 3b crash (which I hadn't seen in awhile), so I decided to try out the other slots in the motherboard (3+4), and when I booted it back up I get 2 long beeps and then a screen saying it failed to boot because of overclocking.

...what is going on? I'm going to put it back to the defaults of 11-11-11-28 and 39 Row Cycle Time and see what that does.

I also have no other computer to put the video card in, not until tomorrow at least, and my motherboard has no integrated graphics so I have to have a video card in.
Should I put my old one in?

 

[koolkat77]

Should I put my old one in?

Yes. How many sticks of RAM do you have exactly?
Sorry if I asked this question earlier

 

[Tezzick]

Should I put my old one in?

Yes. How many sticks of RAM do you have exactly?
Sorry if I asked this question earlier

I have 2. 2x4gb.
And I tried booting it with the RAM in the other 2 slots (1&2) and its booted up just fine.
Seems weird that it won't boot in the other slots, but it will in 1 and 2.

If you'd like, I can give you the exact model numbers for the ram, memory, motherboard, video card, whatever, if that would help?

Edit: Also, forgot to mention, I haven't put my old card in just yet, because the system didn't seem to behave any differently the last time I tried the old one (yesterday).

 

[koolkat77]

I have 2. 2x4gb.
And I tried booting it with the RAM in the other 2 slots (1&2) and its booted up just fine.
Seems weird that it won't boot in the other slots, but it will in 1 and 2.

If you'd like, I can give you the exact model numbers for the ram, memory, motherboard, video card, whatever, if that would help?

Edit: Also, forgot to mention, I haven't put my old card in just yet, because the system didn't seem to behave any differently the last time I tried the old one (yesterday).

That might help Britton30
Did you test the RAM individually or with all inserted?

 

[Tezzick]

I have 2. 2x4gb.
And I tried booting it with the RAM in the other 2 slots (1&2) and its booted up just fine.
Seems weird that it won't boot in the other slots, but it will in 1 and 2.

If you'd like, I can give you the exact model numbers for the ram, memory, motherboard, video card, whatever, if that would help?

Edit: Also, forgot to mention, I haven't put my old card in just yet, because the system didn't seem to behave any differently the last time I tried the old one (yesterday).

That might help Britton30
Did you test the RAM individually or with all inserted?

Haha well if he comes around again, then.
I moved both sets over, I haven't yet tried one at a time, yet.
I can do that in an hour or so.
Is there a specific way I should go about it?
Or just go one stick a time 1 through 4?

 

[koolkat77]

Haha well if he comes around again, then.
I moved both sets over, I haven't yet tried one at a time, yet.
I can do that in an hour or so.
Is there a specific way I should go about it?
Or just go one stick a time 1 through 4?

Yup..1 at a time through 4

 

[Dwarf]

Not being able to boot with memory in some slots, but being able to with it in others, points to a motherboard problem. I'd like you to run Memtest with each stick of RAM in turn in each of the slots. So you'd run it 3 times with the first stick of RAM in Slot 1, then 3 times with the same stick in Slot 2. Repeat for all slots. Then repeat the whole thing with the other stick(s) of RAM. Allow a minimum of 3 passes each time, and let us know what the results are. Do you get memory errors in some slots but not others, regardless of which stick of RAM is inserted? Is/are it the same slot(s) giving the errors each time?

 

[Tezzick]

Not being able to boot with memory in some slots, but being able to with it in others, points to a motherboard problem. I'd like you to run Memtest with each stick of RAM in turn in each of the slots. So you'd run it 3 times with the first stick of RAM in Slot 1, then 3 times with the same stick in Slot 2. Repeat for all slots. Then repeat the whole thing with the other stick(s) of RAM. Allow a minimum of 3 passes each time, and let us know what the results are. Do you get memory errors in some slots but not others, regardless of which stick of RAM is inserted? Is/are it the same slot(s) giving the errors each time?

Alright. I'll start with the first stick overnight, and continue with the rest tomorrow.
Thanks for the help so far guys.

 

[Dwarf]

Just another thought. Since the memory controller is built-in to the CPU, it could be that is faulty and not the motherboard. Have you got another CPU that you could try?

 

[Tezzick]

I do not have another CPU compatible with this motherboard. I only have one computer.
Also update from overnight, the first memory stick ran 11 passes without error in slot one.
However, when I tried testing it in slot 3 (the next slot over on this motherboard), it would not even boot; giving the two long beeps then stating the computer could not boot because of overclocking... Same can be said for slot 4. Slot 1 and 2 work fine.

I do have the ram changed from the auto default of 11-11-11-28 to 9-9-9-24, would that be considered overclocking?
And if so, why does it work in the first slots but not the second slots?

One last thought, I read that AMD Phenom II x4 965 BE's only support 1333mhz ram... and if I wanted any higher speed out of my ram I would have to overclock the CPU first. BUT, when I first booted the board, right after I put it all together over a week ago, it had my ram speed at a default of 2000mhz, not 1333 like many people were claiming.

For the CPU, is there stress-testing that could be done for it?

I almost feel like RMA-ing everything and starting over...

 

[Britton30]

Here is your RAM from the number in your CPUz snip, where this has an 4GBRL, this is 8GBRL which means it is an 8GB kit.
These are the settings it should work at. You can try bumping the RAM volts by .125V or so. If you have CPU-NB volts, raise it by .200V.

View attachment 244665

 

[Tezzick]

Here is your RAM from the number in your CPUz snip, where this has an 4GBRL, this is 8GBRL which means it is an 8GB kit.
These are the settings it should work at. You can try bumping the RAM volts by .125V or so. If you have CPU-NB volts, raise it by .200V.

View attachment 244665

I'll try fiddling with the voltages after some more testing then.

I've been running Prime95 blend test, and after 10 minutes my CPU is at 61 degress Celsius... the max temperature is 65 I'm pretty sure. Is this bad?
EDIT: Also, bear in mind this is with my case opened... I can just put my hand overtop, inside the case, and I can feel it being very hot in there.

Edit2: Oh and is the tRC fine at 33? For the memory.

 

[Tezzick]

Updates: Ran Prime95 blend test for 2 hours with no errors, CPU max temp was 61 but stabilized at around 58 when I turned my case fan to high.
Computer hasn't crashed for 4 hours now, nothing unusual about that, though. Seems to run for a good few hours before crashing.
Currently running Video Memory Stress Test for the next 2 hours. I'll post if that shows anything, and then I'll probably to Furmark afterwords.

Update 2: Completed Video Memory Stress Test and Furmark with no errors, or artifacts, or crashes.
Furmark leveled out at 60 degrees after running for about 15 minutes.

 

[Britton30]

Sounds promising now. Yep tRC should be fine at 33-34, it's not so critical but stay close to 33.

 

[Tezzick]

Sounds promising now. Yep tRC should be fine at 33-34, it's not so critical but stay close to 33.

Ya seems to be running just fine now.
No crashes at all yesterday.
But there is still the matter of the ram not wanting to boot in the other slots. I think I'll just ask Gigabyte about that one.

Still a little too soon to say solved on this just yet, but if it doesn't crash at all in the next few days of use, I'll mark it solved.

 

[Britton30]

Sounds good. You may have a bad RAM slot. If Gigabyte offers to RMA it for you see if they have an option to get a replacement then send your old one in. That way you will get another board and not your old one back.

 

[Tezzick]

Sounds good. You may have a bad RAM slot. If Gigabyte offers to RMA it for you see if they have an option to get a replacement then send your old one in. That way you will get another board and not your old one back.

Oh alright. They'd normally try to fix it and send it back vs a new one?

Also, 38 is a good temp for a CPU with minor load? (just internet browsing and wayching live streams)
My case is pretty old (Antec Sonata III), with only a rear exhaust fan and I was thinking of cutting out a spot in the top of it for another 120mm fan to possibly increase air circulation. Not exactly sure if needed though? Those temps are with the case open, and on its side, though. It goes up another 3-5 degrees closed and upright.