Help diagnosing Win7 crashing problem

tim wrote in post #12406187
Thanks for the Blue Screen software link krb. Below is what it says... basically it's the NT kernel crashing, not a driver.

Could an overheating video card do this? A quick google suggests 74 is within the operating range of the card.

051411-11980-01.dmp 14/05/2011 2:30:29 p.m. SYSTEM_SERVICE_EX​CEPTION 0x0000003b] 00000000`c0000005 fffff800`02afd431 fff​ff880`0847bd00 0000000​0`00000000 ntoskrnl.ex​e ntoskrnl.exe+7fd00 N​T Kernel & System Microsoft® Windows® Operating System Microsoft Corporation 6.1.7601.1​7592 (win7sp1_gdr.110408-1631) x64 C:\Windows\​Minidump\051411-11980-01.dmp 4 15 7601 291,8​72

FYI - The x05 code is a memory access exception. One or another of your processes apparently attempted to write or branch to an invalid (protected) location in the kernel. If you search the minidump and find the process name win32k.sys, you might be suspicious of your video card. It's one of the few user drivers that are allowed to link into ntos territory.

The PC has been totally stable since I took the extra RAM out, so i'm going to blame that. I've had 3 days uptime over the course of a week of us, including a bunch of standby/resume cycles, so I think it's all good now.

Thanks to everyone who helped track the problem down

Unless you have an IR thermometer, you're relying on a tiny thermistor reading to tell you the core temperature? I'd pull the cooler off and make sure it is seating correctly on the GPU, check whether there is enough heat transfer paste on there, better yet just throw an air cooler on there.

Frank_The_Tank wrote in post #12583898
Unless you have an IR thermometer, you're relying on a tiny thermistor reading to tell you the core temperature? I'd pull the cooler off and make sure it is seating correctly on the GPU, check whether there is enough heat transfer paste on there, better yet just throw an air cooler on there.

Thanks for your thoughts Frank, but problem solved - it was mismatched RAM.

tim wrote in post #12583956
Thanks for your thoughts Frank, but problem solved - it was mismatched RAM.

go on - put them back in to be sure...

I could probably put the other ones back in and have them work fine, maybe. I'll just put them on trademe.

tim wrote in post #12583527
The PC has been totally stable since I took the extra RAM out, so i'm going to blame that. I've had 3 days uptime over the course of a week of us, including a bunch of standby/resume cycles, so I think it's all good now.

Thanks to everyone who helped track the problem down

Well, good luck. Just a note, though. This thread is interesting because I have two almost identical machines - same MB, same OS, same size and brand drives. One has 8GB, the other has six, the first is a 9550 and the second a 9400. The the 8MB machine is matched RAM, the six MB is unmatched.

The 8MB keeps failing on memory exceptions, or execution exceptions in low ram at random intervals. If I take out either bank of RAM down to 4GB, it runs forever. The 6GB machine has never had the problem.

Curious, huh?

Hen3Ry wrote in post #12585365
Well, good luck. Just a note, though. This thread is interesting because I have two almost identical machines - same MB, same OS, same size and brand drives. One has 8GB, the other has six, the first is a 9550 and the second a 9400. The the 8MB machine is matched RAM, the six MB is unmatched.

The 8MB keeps failing on memory exceptions, or execution exceptions in low ram at random intervals. If I take out either bank of RAM down to 4GB, it runs forever. The 6GB machine has never had the problem.

Curious, huh?

This is pretty common. Sometimes increasing the voltage to your memory will resolve the issue but do so at your own risk.