Sign in with
Sign up | Sign in
Your question
Solved

Mem Management, IRQL, Page fault BSODS + Memtest Errors

Last response: in Memory
Share
October 2, 2012 4:00:58 PM

Homebuilt PC, running 8gb (2x 4gb) ram sticks; Crucial Ballistix sport 4GB 240-Pin DDR3 SDRAM DDR3 1333 (PC3 10600) Desktop Memory Model BLS4G3D1339DS1S00

Mobo is msi g45 870.

Got the errors on memtest while running with both of them so now i'm running seperatly to see which is the problem.

but my questions is how do i know if its the ram, ram settings, or the mobo? Right now im running the ram sticks individually to see if they get any errors, but lets say hypothetically there are no errors, what then? or lets say there are errors, what do i do? how do i rule out that it's not the mobo settings.... not overclocked. Running a coolermaster 500w psu just in case that matters. Thanks for any help.


running sticks separately now.. first test 9 passes no errors but as soon as i go to shutdown the PC i get a IRQL BSOD so i don't know what to make of it....

2nd ram stick ran for 4hours, 4 passes, no errors.......... what am i suppose to make of this?
a c 347 } Memory
October 2, 2012 4:19:06 PM

Welcome to Tom's Forum! :) 

A "IRQL BSOD" can be from a bad application to a bad driver, so you'll need to look at the 'Event Manager' to begin to reveal the core issue; look for [+] Critical errors - http://i1013.photobucket.com/albums/af254/Jaquith/Event...
* Run MSCONFIG and Diagnostic Mode and see if the BSOD is corrected. Memory errors clearly can induce a BSOD.

Memory errors - if each stick by itself passes individually but fails together then this can be a symptom of:
* Bad BIOS (often old especially on AMD); solution update the BIOS.
* Bad DIMM slot(s); sometimes an after market HSF with metal parts including the back plate can cause a grounding or short a DIMM slot, or incorrect DIMM slots used; use the two Black DIMM slots.
* Kit vs Non-Kit - this is often a problem with higher frequency RAM, but 'generally' no so much with DDR3-1333 sticks.
* Bad stick of RAM.
* Older version of Memtest; use the latest v4.20 - http://www.memtest.org/#downiso

MSI 870-G45 - http://www.msi.com/product/mb/870-G45.html
Crucial BLS4G3D1339DS1S00 - http://www.crucial.com/store/partspecs.aspx?IMODULE=BLS...

Q - What CPU?
Q - What BIOS version? ; you can use CPU-z (Mainboard tab/BIOS section) to identify; link - http://www.cpuid.com/softwares/cpu-z.html
m
0
l
October 2, 2012 4:24:27 PM

im using a amd phenom x4 955 cpu, bios version is a bit old and going to flash but i know you're not supposed to flash unless the system is stable.. since its unstable i dont want to flash the bios yet...

Was using a ballistix kit that i purchased and i am using the latest memtest version.

looking at the event viewer each BSOD says "Kernel-Power" bluescreen viewer says ntoskrnl.exe for both mem management and irql bsods...

i hope by kernel power it doesnt mean literal power. meaning my power supply =\

and thanks for the welcome :D 
m
0
l
Related resources
a c 347 } Memory
October 2, 2012 4:51:27 PM

More than likely this is a RAM, memory, based issue. In other words that type of failure is H/W related.

Quote:
running sticks separately now.. first test 9 passes no errors but as soon as i go to shutdown the PC i get a IRQL BSOD so i don't know what to make of it....


Those are two separate environments, Memtest is non-OS and an IRQL 'BSOD' is OS related.

Therefore, if the PC is stable in the BIOS and Memtest passes then you're safe to update the BIOS, but IF when you're in the 'BIOS' and the PC shuts down unexpectedly then be clear don't attempt to update the BIOS.

The only safe way to update a BIOS is 'in' the BIOS, and MSI offers M-Flash. Here's a lack luster YouTube on M-Flash but you get the idea - http://www.youtube.com/watch?v=H7ws0deuHL4
m
0
l
October 2, 2012 4:54:51 PM

jaquith said:
More than likely this is a RAM, memory, based issue. In other words that type of failure is H/W related.

Quote:
running sticks separately now.. first test 9 passes no errors but as soon as i go to shutdown the PC i get a IRQL BSOD so i don't know what to make of it....


Those are two separate environments, Memtest is non-OS and an IRQL 'BSOD' is OS related.

Therefore, if the PC is stable in the BIOS and Memtest passes then you're safe to update the BIOS, but IF when you're in the 'BIOS' and the PC shuts down unexpectedly then be clear don't attempt to update the BIOS.

The only safe way to update a BIOS is 'in' the BIOS, and MSI offers M-Flash. Here's a lack luster YouTube on M-Flash but you get the idea - http://www.youtube.com/watch?v=H7ws0deuHL4



i was going to use the usb flashing tool to update the bios. i got the bsod after running the memtest then booting into windows then shutting down as the pc was shutting down i got the BSOD...

but if its a hardware (memory) issue why is it that they dont fail individually? any idea of how to remedy the situation or do you think the remedy is to update the bios?
m
0
l
a c 347 } Memory
October 2, 2012 5:07:02 PM

magepker said:
i was going to use the usb flashing tool to update the bios. i got the bsod after running the memtest then booting into windows then shutting down as the pc was shutting down i got the BSOD...

but if its a hardware (memory) issue why is it that they dont fail individually? any idea of how to remedy the situation or do you think the remedy is to update the bios?

Simple, it's a sign of a bad BIOS. Often with AMD I see 4GB/stick memory issues especially with older BIOS versions, and the BIOS + OS aren't properly communicating.

You can first try to properly clear the BIOS, most folks do it wrong. Unplug the PSU for 5 minutes (discharge the PSU), move the CMOS Jumper for 5+ seconds to the 'Clear Data' XXO -> OXX (5+ seconds) -> XXO, plug the PSU back in, boot to BIOS, (F6) Load Optimized Defaults = OK, and (F10) Save & Exit = OK. The see what happens. Nice video - http://www.youtube.com/watch?v=rdHH9KrceR0

When to update a BIOS:
1. CPU requires newer BIOS version
2. Non-OS Instability
3. Peripherals e.g. 2TB HDD requires new BIOS.
4. Memory compatibility ; where current RAM is causing errors or stability.
m
0
l
October 2, 2012 5:10:21 PM

alright well, just updated my bios to the newest version, surprised that its almost 4 versions old.... cleared the cmos, loaded optimized settings (date\time was messed up so i fixed that)

Hopefully that fixes it... wont know until i get another BSOD... but in general i've been getting really lackluster computer performance but i have no way of finding out what the problem is.... i used easy driver pro to update my drivers but i'm starting to think that is the problem, maybe it downloaded and installed the wrong drivers but how can i even check?

i started driver verifier, hopefully that comes up with something on the next BSOD if i ever get one.

thanks for the help btw.

edit: i did some research and saw that it may also be a driver issue, so i downloaded driver sweeper and it states i have nvidia chipsets.... my board is an AMD 770 chipset running an ATI card... im confused...
m
0
l
a c 347 } Memory
October 2, 2012 5:28:07 PM

Well the only (3) places most 'should' get their drivers from are:
1. MSI for your specific MOBO and OS including bit size; e.g. Windows 64-bit - http://www.msi.com/product/mb/870-G45.html#/?div=Driver...
2. Use Windows Update
3. Discrete GPU's:
a. AMD - http://support.amd.com/us/Pages/AMDSupportHub.aspx
/or
b. nVidia - http://www.geforce.com/drivers

Going to an OEM site or worst a 'third party application' can end up causing a disaster (drivers with ROM/Firmware) or render the OS with severe driver and OS corruption and or conflicts.
m
0
l
October 2, 2012 5:30:05 PM

well i just realized that they're drivers installed directly from windows 7.

so do you think this BIOS update will fix the bsod issues?
m
0
l
a c 347 } Memory
October 2, 2012 5:31:04 PM

Driver sweeper was taken off the market because it misidentifies drivers and removes the wrong stuff.

My best advice is to re-install the #1~#3 links above and re-install all of the drivers from ONLY the correct sources.
m
0
l
a c 347 } Memory
October 2, 2012 5:34:30 PM

magepker said:
so do you think this BIOS update will fix the bsod issues?

Time will tell.

I would look into the drivers if you have any more issues and re-install the 'correct' drivers only.

You can also stress test your PC using AIDA64 Extreme and select all stress testing options as follows, and lastly re-test Memtest. My general advice is AIDA64 Extreme for 24 hours and Memtest either overnight or a minimum of 4 Full Passes.

AIDA64 Extreme - http://www.aida64.com/downloads

m
0
l
October 2, 2012 5:45:27 PM

alright thanks a lot jaquith really appreciate it. ill definitely run the stress test, last question, should i run memtest with the sticks individually again? or in dual channel since thats how i got the errors in the first place?

and what steps should i take if i get errors again...
m
0
l
a c 347 } Memory
October 2, 2012 6:00:55 PM

Run Memtest with all RAM in the correct DIMM slots - use the two Black DIMM slots.

IF you get errors then in the BIOS try to increase both the DRAM Voltage -> 1.55v (or +0.05v) and CPU-NB Voltage -> 1.15v (or +0.05v).
m
0
l
October 2, 2012 7:52:03 PM

sigh.... after an hour of memtest with both ram sticks, got errors... increased the dram voltage like you said, went to increase cpunb voltage to 1.15 but auto is already @ 1.65.... so i put it + 1 notch so its @ 1.69 now. going to run the stress test then memtest again.

ran aida.... hardware failure detected within like 30seconds....
m
0
l
a c 347 } Memory
October 2, 2012 8:18:26 PM

magepker said:
sigh.... after an hour of memtest with both ram sticks, got errors... increased the dram voltage like you said, went to increase cpunb voltage to 1.15 but auto is already @ 1.65.... so i put it + 1 notch so its @ 1.69 now.

Using the 'correct' CPU-NB Voltage -> 1.10v is 'normal' and +0.05v helps and +0.10v helps with DDR3-1600. My guess is you're looking at the NB Voltage.
m
0
l
October 2, 2012 8:27:25 PM

well like i said under cpu nb auto, says min is 1.61 auto is 1.65.... i can post a picture if that would be any help... is there a way to check the voltage without having to get into bios each time?


edit: after getting back into the bios i noticed i've been giving you the wrong #'s.

Under CPU-NB says auto is 1.265V so i upped it to 1.269V which is 1 notch.

But i noticed theres an option that says just NB which is auto @ 1.103 is that the one i'm supposed to change?
m
0
l
a c 347 } Memory
October 2, 2012 8:58:07 PM

magepker said:
well like i said under cpu nb auto, says min is 1.61 auto is 1.65.... i can post a picture if that would be any help... is there a way to check the voltage without having to get into bios each time?


edit: after getting back into the bios i noticed i've been giving you the wrong #'s.

Under CPU-NB says auto is 1.265V so i upped it to 1.269V which is 1 notch.

But i noticed theres an option that says just NB which is auto @ 1.103 is that the one i'm supposed to change?

I was referring to your 'manual' page 3-23, and the correct 'CPU-NB Voltage' should be around 1.10v (sure there's some variations) -- so "NB which is auto @ 1.103" is probably 'it' and up that to 1.15v.

Now if for some reason that fails, no doubt WE hope it'll pass, then try the Blue DIMM slots and run Memtest again.
m
0
l
October 2, 2012 8:59:59 PM

ah ok, sorry for the confusion. i'll change it and run memtest again.
m
0
l
October 2, 2012 9:05:45 PM

Trying change the default nb voltage automatically changes the sb and ht link voltage, not sure thats normal or not
m
0
l

Best solution

a c 347 } Memory
October 2, 2012 10:36:26 PM

Just leave it set and +0.05v is very minimal.
Share
October 3, 2012 12:33:23 AM

alrighty. so after about 3 and a half hours and 3 passes no errors! looks like it did the trick since the first 2 times i got errors within the first 2hours!

thanks for your help jaquith, if anything happens again ill just repost here >_<
m
0
l
October 3, 2012 2:01:41 PM

Best answer selected by magepker.
m
0
l
a c 347 } Memory
October 3, 2012 3:13:30 PM

Very Cool! :) 
m
0
l
October 5, 2012 1:52:37 AM

*** on a stick, 2 days of no BSOD's and i just get one.. MEMORY MANAGEMENT, just no clue what to make of it anymore. any suggestions?
m
0
l
a c 347 } Memory
October 5, 2012 1:31:57 PM

In order to get DDR3-1600 to run with a non-FX CPU requires you to OC the FSB from 200MHz to 240MHz which in turn OC's the CPU so to undo that OC you then need to lower the CPU Multiplier. Further, the performance difference is negligible with the faster frequency RAM, so unless the goal is to also OC the CPU I really don't recommend faster than DDR3-1333.
m
0
l
October 5, 2012 3:03:12 PM

oh i see... so ill just buy the equivalent ddr3 1333.. same price anyway lol.
m
0
l
a c 347 } Memory
October 5, 2012 3:37:32 PM

DDR3-1600 kits will Default to DDR3-1333 so all I'm suggesting is that you realize that upfront. Also, if you're buying RAM then get a 2x4GB kit. However, since you have a working 2x4GB kit now I see no reason to replace your current RAM. Most RAM carries a Lifetime Warranty...
m
0
l
October 5, 2012 3:42:25 PM

well the 2x4gb kit isnt working since its giving me BSOD's >_<
m
0
l
a c 347 } Memory
October 5, 2012 4:06:11 PM

Oh?!
Quote:
alrighty. so after about 3 and a half hours and 3 passes no errors! looks like it did the trick since the first 2 times i got errors within the first 2hours!

CPU-NB Voltage -> 1.20v ; +0.10v ; still safe max safe is 1.30v
DRAM Voltage -> 1.55v~1.60v ; +0.05~+0.10v ; still safe max safe is +0.15v~+0.20v
m
0
l
October 5, 2012 9:04:07 PM

done. hopefully this fixes it >_<
m
0
l
a c 347 } Memory
October 5, 2012 9:13:37 PM

Well...you know what to do next.

Assuming it all passes, it's probably going to be a good idea to run SFC and I'd schedule both options especially for your primary OS (C: ) disk.

m
0
l
October 5, 2012 9:15:17 PM

i actually changed those values yesterday, ran aida overnight (12ish hours) nothing bad.... booted into memtest for about 8hours (3 passes) no errors. ill run a chkdsk. thanks.
m
0
l
a c 347 } Memory
October 5, 2012 9:19:47 PM

Very Cool (take 2)! :) 
m
0
l
October 5, 2012 9:23:39 PM

yep :D , hopefully i never have to talk to you again lol.
m
0
l
a c 347 } Memory
October 5, 2012 9:36:25 PM

Well at least for 'this' reason -- Good Luck! :) 
m
0
l
October 11, 2012 8:22:32 PM

sigh... 6 days later just got a BSOD... i dont think upping voltages will help again, do u think i should just buy new ram?
m
0
l
a c 347 } Memory
October 11, 2012 9:09:55 PM

What's the error code for the BSOD?

At some point as I've alluded to earlier is a match kit of RAM vs a couple individual sticks; Kit vs Non-Kit.
m
0
l
October 11, 2012 9:16:08 PM

well i am using kits... i bought 2 seperate kits, 1 is 2 x4gb and the other was 2x 2gb...

error code was page fault in non paged area pointing to ntoskrnl.exe again

101112-26301-01.dmp 10/11/2012 4:19:26 PM PAGE_FAULT_IN_NONPAGED_AREA 0x00000050 ffffba80`0c9ba938 00000000`00000000 fffff800`03378a5e 00000000`00000007 ntoskrnl.exe ntoskrnl.exe+7efc0 NT Kernel & System Microsoft® Windows® Operating System Microsoft Corporation 6.1.7601.17944 (win7sp1_gdr.120830-0333) x64 ntoskrnl.exe+7efc0 C:\Windows\Minidump\101112-26301-01.dmp 4 15 7601 275,304

thats what bluescreenview says.
m
0
l
a c 347 } Memory
October 11, 2012 9:39:13 PM

Okay, if you're running 2x4GB + 2x2GB together then that sure can be a problem. For starters, CAS Timings, maybe ranking, etc and hopefully they're installed: CPU: | 2GB | 2GB | 4GB | 4GB | or | 4GB | 4GB | 2GB | 2GB |.

The BLS4G3D1339DS1S00 is the individual stick whereas the BLS2KIT4G3D1339DS1S0 is the 2x4GB kit.

Please provide links to each set of RAM or CPU-z 'SPD' tab information (screen shots) for each density 2GB & 4GB.

There is a specific Hotfix on the OS level that 'might' be the cause, see - http://support.microsoft.com/kb/951418 other than that I'd need to the actual Dump file, and on my new rig I don't want to install any software to read those files. These guy's however will take a look - http://www.sevenforums.com/
m
0
l
October 11, 2012 10:56:03 PM

i get bsods when running 12gb of ram but i also got the bsods when running on the 2 x 4gb sticks alone.. no bsods when running just the 2x 2gb sticks. i bought the kits @ microcenter but pulled the product info from the crucial website... by the looks of cpu-z guess i got shafted, guess they just put the individual sticks together and called them a kit smh....

heres the picture u requested;
m
0
l
October 11, 2012 11:13:54 PM

btw tried the hotfix, for some reason says it doesnt apply to my computer.
m
0
l
a c 347 } Memory
October 12, 2012 4:16:21 PM

magepker said:
i get bsods when running 12gb of ram but i also got the bsods when running on the 2 x 4gb sticks alone.. no bsods when running just the 2x 2gb sticks.

Seems like a a bad 4GB stick so my advice is to contact Crucial and RMA both 4GB sticks. IF the RAM was shipped in ONE box with TWO sticks then they sold you a set.

Memtest proves the RAM does or doesn't throw errors only for the small window of time while being tested. All RAM throws 1-2 errors per month the only difference is ECC fixes it and non-ECC doesn't. Depending on both the number of errors and what's the error a bad stick often causes a seeming random BSOD.

I assume you're using the latest BIOS.
m
0
l
October 12, 2012 5:25:33 PM

yea updated bios late september/early october.... was trying to avoid RMA but guess thats the only thing to do,

btw would u want my to post the txt versions of the dumps?
m
0
l
a c 347 } Memory
October 12, 2012 6:03:20 PM

If you click on the 'Event Log Online Help' often there's some useful information. Critical Errors are the ones that really matter the most. Particularly if they're frequent, the same and persistent.

My way of dealing with persistent OS issues is to re-install the OS and not trying to fix it.

These folks are your best bet - http://www.sevenforums.com/ however if the 2x2GB RAM solves all of the issues then it seems it's the RAM and the voltages don't seem to correct the problem.
m
0
l
October 20, 2012 8:15:52 PM

jaquith said:
If you click on the 'Event Log Online Help' often there's some useful information. Critical Errors are the ones that really matter the most. Particularly if they're frequent, the same and persistent.

My way of dealing with persistent OS issues is to re-install the OS and not trying to fix it.

These folks are your best bet - http://www.sevenforums.com/ however if the 2x2GB RAM solves all of the issues then it seems it's the RAM and the voltages don't seem to correct the problem.


i waited a week before i posted again, last BSOD was on the 11th. been 9 days since my last BSOD and seems like upping the voltage again did the trick. i hope.

i did as you said and posted on sevenforums and they were stumped and told me to just RMA the ram, to me that was the absolute last step so i didnt do it...
m
0
l
a c 347 } Memory
October 20, 2012 9:07:27 PM

Well I'm sort of in agreement with then, the RAM carries a Lifetime warranty and you have other RAM in the mean time until shipping turnaround is done...
m
0
l
!