• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Need Some Help - Not Sure if it's Memory, CPU, or Motherboard

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

JrClocker

AKA: JrMiyagi
Joined
Sep 25, 2015
I am reaching out to the big brains here at OverClockers.com...I need some help.

History:
- Off and on for the past few months, I have been having random PC crashes on "The Sith" below. The machine has 2 GPUs and runs 24/7. The random crashes would only occur while running folding@home.
- These crashes would reset the PC, and I would be presented with the Windows log in screen.
- This weekend, it crashed every day.
- The big difference yesterday, is that it crashed while I was flying my spaceship in Elite Dangerous.

OK - so I can't blame folding@home anymore...

I told Windows to run a memory check. It found an error within the first 5 minutes.

OK - so it's a memory problem I say.

I play the memory stick swap out game, and still get the error within the first 5 minutes.

I drop my modest memory overclock from 2666 MHz, to 2400 MHz, to 2133 MHz. Memory error within first 5 minutes.

I start some Google-fu...change my SA voltage from Auto to 0.95.

No memory errors at 2133, 2400, or 2666 MHz.

Everything boots into Windows fine. Within 5 to 10 minutes of logging in, I now get consistent BSOD BAD_POOL_CALLER.

I bring all overclocks back to stock (CPU, Memory, etc.)

Currently running on 16 GB of RAM (didn't put the other 4 sticks back in).

Still getting BSOD BAD_POOL_CALLER within 5 to 10 minutes of logging in (even if I just let the machine sit there and do nothing.)


The hardware is all relatively new, so I don't think I have a bad PSU.


Things I am going to try:
- Add the other 4 sticks of RAM in to bring to 32 GB. Maybe having 2 12 GB video cards needs more than 16 GB of RAM?
- Update MB BIOS to latest (I hate updating BIOS). Maybe some of the Windows Specter/etc. patches are causing an issue.
- Remove my Sound Blaster PCI audio card (the last driver update was 12/2016).
- Go back to last restore point.
- Boot into Safe Mode and reinstall all drivers (uggg).
- Lastly...reinstall Windows (not looking forward to this.)


Any other thoughts on what to try?
 
At places like this, I usually boot from a USB stick to another OS. I even have a "Windows to Go" stick I use with the same version of windows I run normally on it. If I am good on the USB boot, testing the same thing that caused the BSOD, then it has something to do with the OS most likely. First thing I do then is check for bad signature files using sfc /scannow in an elevated command prompt. Hopefully that's good because it can be a brute to fix if the auto repair does not do it. Then I do a full system scan for virus. I do that in safe mode, or I can use a YUMI USB stick and run one of the many antivirus scantools I keep on it. Since your getting a BSOD, you can also search the web using the error code related to the BSOD. Best guess is, you have some corrupted system files that will have to be fixed.

-Rodger
 
Added in the 4 memory sticks bringing to 32 GB...BZZT

Made a Windows 10 boot FLASH drive and did "fix it Windows"...BZZT

Went to my last restore point (3/11/2018)...I upgraded to the latest version of Corsair Link...DING!

System has been stable at the desktop for 8 minutes now...typing this from the PC...

Either the latest Corsair Link was bogus...or the memory issues I was having before I increased SA voltage caused an issue with the install. Corsair Link connects to my Corsair H80i v2 cooler through USB.


/crosses_fingers
 
Added in the 4 memory sticks bringing to 32 GB...BZZT

Made a Windows 10 boot FLASH drive and did "fix it Windows"...BZZT

Went to my last restore point (3/11/2018)...I upgraded to the latest version of Corsair Link...DING!

System has been stable at the desktop for 8 minutes now...typing this from the PC...

Either the latest Corsair Link was bogus...or the memory issues I was having before I increased SA voltage caused an issue with the install. Corsair Link connects to my Corsair H80i v2 cooler through USB.


/crosses_fingers

My H80i v2 is sometimes acting weird. All works fine but it randomly sets fan speed to max even though in profiles all is locked at low %. I'm not sure if it should even affect OS stability in any way. I mean it's only cooler + fan/LED management. Software has delayed start.
Since you are not posting more then I guess it works fine now.
 
What was the SA voltage in HWiNFO at before and after you increased it?

Don't know what the SA voltage was before. It was on Auto, and the system wasn't stable enough for me to even load HWiNFO.

After I did the role back to the system restore point mentioned above, I had one more BSOD.

The only other update I did recently was graphics driver update. So, I went into Safe Mode, ran DDU, and reinstalled the graphics driver. Just to be safe, I also reinstalled all of the motherboard related drivers.

Additionally, to be safe, I updated the system BIOS to the latest revision.

System has been running for 36 hours now with no issues (it's folding again...yay!)

The other interesting twist from last night is that the center/sub channel on my SoundBlaster Z sound card stopped working. Sound would not come out for either analog or optical connections. I thought it might be my surround sound system (it's my old one I bought back in 1998 or so).

To be "sure", I connected the sound to the motherboard sound. Center/Sub channel was working again. (The SoundBlaster Z sound is SOOOO much better).

Maybe the sound card driver caused all this...don't know.

I ordered a replacement sound card (the new SoundBlaster with integrated RGB LED controller...RAWR!)
 
Had a couple more BSD this weekend.

I was able to narrow it down to bad memory.

I have 8 sticks of 4 GB. Narrowed it down to "at least 1 of these 4 is bad". Running at 16 GB currently (the minimum recommended for 2 Titan X Pascal video cards.) System has been running now for 48+ hours.

New memory arrives today (Corsair RGB 4 x 8 GB, 3600 MHz).
 
Last edited:
HJopefully that solves things. Picking out random errors can take a while
 
How did you test the memory that is found defective? I hope you have good luck.:)

1.) Test to see if dual channel memory mode works
---> Plug only 2 of 8 memory modules in. Run Windows Memory diagnostic
---> Got lucky and the 2 modules I chose passed the test (i.e. "good")

2.) Test to see if triple channel memory mode works
---> Plug 1 more memory module in
---> Failed the Windows memory test...set it aside
---> Plugged a different memory module in...test passed

3.) Test to see if quad channel memory mode works
---> Plugged "bad" module from step (2) in...verify that it fails memory test (it did)
---> Plug a different memory module in...test passed
---> I now have 4 "good" sticks of 4 GB each for 16 GB

4.) Test to see "double" quad channel memory mode works
---> Plugged all 8 modules in...test fail
---> Removed all 4 added modules to repeat step 3...test passed


So...at least 1 of the 4 sticks is bad. I have 4 good sticks giving me 16 GB.

4 sticks at 8 GB each arrive today to get me back to 32 GB!
 
Thanks!

I got lucky finding 2 "good" modules so quickly...finding a "bad" module soon after that...and getting 2 more "good" modules soon after that!

The new memory has been delivered to my house already...just waiting for the work day to end to play with my new RGB LED pretties!!!!!
 
It’s very pretty!

Have it running at 2,666 MHz...haven’t done any major tweaking. Passes Windows memory test.

It’s so pretty!
 
Back