Win7 64-bit BSODs... please help

[mhowie]

I encountered a period of BSODs nearly one year ago and was helped by the experts in this forum and would appreciate your guidance again. Same machine, rock solid for nearly a year, but has begun BSOD'ing again the past week.

I returned from a business trip last week and upon powering on my PC noticed a weird screen...my computer had booted into its D: drive and found my old XP build. I went into the BIOS and the boot sequence had changed its order (weird). I changed it back, did a little research, and determined my CMOS battery might need swapped out. Soon thereafter I installed the new IE9 RC and began getting BSODs. I wound up uninstalling IE9 RC, rolling back to IE8, and all was well until this evening.

I swapped out the CMOS battery and restarted my rig. All was well. I then decided to give IE9 RC another shot so I redownloaded and installed. Yep, BSOD soon thereafter. Unfortunately, this time, upon uninstalling and reverting back to IE8, the BSODs continue.

I am attaching the Minidumps from tonight and would appreciate your review and subsequent counsel/direction. Perhaps the whole IE9 RC correlation is oddly coincidental and I have some hardware item going bad...but before I commence that level of troubleshooting I thought there would be benefit in understanding what the Minidumps reveal.

Unfortunately as is often the case with the Windows System Restore function, mine is flaky and never completes a restore when attempted...so I can not simply roll back to an earlier time.

Thanks in advance for your help,

Howie

 

[mhowie]

I'm also including those Minidumps from last weekend in case they reveal something-- this is when the BSODs began as noted above.

 

[mhowie]

Any thoughts on what might be the culprit(s)?

 

[Mark56]

The prime suspect here is faulty RAM. Run through this guide and post back the results.

Power off the PC and remove the mains connection, always touch the metal of the case with your hands to avoid any static electricity from causing component damage. (On laptops ALWAYS remove the battery before removing or replacing memory modules). Remove all your memory modules. Clean the contacts with a soft rubber (up and down not from end to end) then insert just one module and reboot to test for any change in performance. Repeat this for every module one at a time. Obviously if you only have one module this cannot be done, but you should still remove and clean the module and reseat it and test.

If you only have one stick download this to test it. http://www.memtest.org/

You need to scroll down the page until you see the three blue screenshots. Just below is the list of downloads for Memtest86+ V4.10. Click on Download - Pre-Compiled Bootable ISO (.zip)and save to your desktop, When the download is complete right click the file and select Extract Here and burn the image to a CD. In windows 7 right click the extracted file, select Open With, then select Windows Disc Image Burning Tool then follow the prompts. For all other versions of windows (if you do not have an ISO burner) download this free software. http://www.isoimageburner.com/

Boot the PC into the Bios setup and set the CD/DVD drive to 1st in the boot sequence. Insert the disk in the drive then reboot and the disc will load into dos. Leave the test to run through at least 5 cycles.

The memtest will not be 100% accurate but should easily detect any major faults.

 

[mhowie]

Thanks for the feedback. I have the Windows-based program MemTest...is this program inadequate to test RAM?

 

[Mark56]

As I said in my guide above

The memtest will not be 100% accurate but should easily detect any major faults

And that is why I recommend testing by using just one memory stick at a time and only using the software if you only have one stick.

Memtest for windows willl find major faults but running it off a bootable CD is far better.

 

[mhowie]

This might be a bit premature...but I'm feeling optimistic so I will share details as of this evening.

I started testing with MemTest early this afternoon and immediately received memory error messages-- multiple, within just a few seconds...and the messages were cryptic (e.g., "Module 147132 could not swap with Module 345984"...or things to that effect).

I immediately shut down my PC and removed the RAM. Here is where things may have begun to make sense (assuming this was the source of my problems--again, I'm practicing optimism as I haven't had any crashes in several hours). My motherboard (MSI K9AG Neo2-Digital) is dual-channel capable. Four slots with DIMM1 and DIMM3 comprising Channel A and DIMM2 and DIMM4 Channel B. As I removed the RAM I noted the existing configuration:

DIMM2-- 2GB
DIMM1-- 1GB
DIMM4-- 2GB
DIMM3-- 1GB

As I tried to make sense of the situation, I realized I may not have had them (Corsair RAM...two matching pairs...TWIN2X4096-6400C4DHX and TWIN2X2048-6400C4DHX) properly populated-- I had a matching pair of 1GB sticks in Channel A and the matching pair of 2GB sticks in Channel B. Oddly, CPU-Z had always reported they were Dual Channel and I had been running that orientation for over one year.

So, I removed the 1GB sticks, populated DIMM1 and DIMM2 with the 2GB sticks, and ran MemTest again. This time...no errors after a couple of hours. Throwing caution to the wind, I installed the 1GB sticks resulting in 3GB per Channel:

DIMM2-- 2GB
DIMM1-- 2GB
DIMM4-- 1GB
DIMM3-- 1GB

Started things up, ran MemTest for a couple of hours without any errors, and have been working in multiple applications trying to replicate the crashes of earlier.

Assuming all is well again, I would like to ask the following:

Could RAM orientation as I had it result in the crashes (and associated crash data specifics) experienced?
Could these kinds of problems manifest themselves suddenly after the configuration performing ok for over a year?
Could this funkiness cause the problems with the boot sequence settings being "forgotten"?
Is the stability of the past few hours merely luck and the shifting of RAM sticks as I performed not something that would have such an impact or be particularly relevant?
Anything else I should be on the lookout for?

Appreciate the wisdom...always trying to learn.

Best,

 

[Mark56]

As per my guide the first thing I suggest is reseating the sticks, You may have just had a bad connection.

You would have to check with your motherboard manual but all that I have seen before will run OK with uneven RAM channels, although, it is always recommended that the channels are even and the sticks are matched pairs. Your PC has run without fault until recently with a bad setup so I guess this proves that it was OK.

As this arrangement of RAM sticks has run succesfully for such a long time I believe that one of the sticks is faulty. Balancing the channels, coupled with the limitations of running memtest in windows is most likely covering up the fault.

I would run the PC on each individual stick and then play a game or watch a movie to push the RAM use up high and see if you get any more crashes.

If you want to do a quick test then I would download and burn to disc an up to date copy of Memtest86+ and run it by booting from the disc. This gives it full access to the RAM which is not the case when running it within windows. Let the test go through at least five passes per stick.

In answer to your questions.

Could RAM orientation as I had it result in the crashes (and associated crash data specifics) experienced?
I doubt it very much as it had been running without fault for such a long period.

Could these kinds of problems manifest themselves suddenly after the configuration performing ok for over a year?
Extremely unlikely.

Could this funkiness cause the problems with the boot sequence settings being "forgotten"?
No, that is always caused by a flat battery.

Is the stability of the past few hours merely luck and the shifting of RAM sticks as I performed not something that would have such an impact or be particularly relevant?
This could just be luck, or one or more of the sticks had a bad contact which would have been cured by reseating. If it was not due to a bad contact I believe the errors will happen again.

Anything else I should be on the lookout for?
Nothing I can think of for now.

 

[mhowie]

Thanks again for the wisdom and direction.

I ran MemTest again yesterday for 12+ hours (600%+ coverage each with two instances of it running) and received no errors.

Hopefully the bad contact theory applies...perhaps somehow one of the sticks had become slightly unseated over time? If I do experience these crashes again in the future and have memory errors via MemTest, I will yank the 1GB sticks and just run 4GB RAM until such time as I build a new rig.

Interesting note on the boot sequence settings being forgotten...this occurred one time after I had changed the battery but before this recent stretch of stability.

 

[Mark56]

Glad to hear everything is running smoothly again.

The Bios will always reset itself to default settings when a CMOS battery fails or is removed.