BSOD when GPU overheats... Seems too cool to be overheating.

[Siesna]

My computer fairly consistently BSODs while playing games for extended periods of time. Usually while playing a game my GPU peaks at about 70~72 C. I know this is up there, but I've seen lots of people not having to worry about overheating until they get to 90 or so.

So my question is, is it normal for my computer to be overheating with a GPU at 72? Or could there be something else underfoot that is causing my issue?

Unfortunately, the last few bluescreens I had were too quick to save any minidumps, and all my old ones are gone... I think I may have accidentally wiped them with CCleaner or something. I've been having this problem for a while and have finally (I think) narrowed it down to my GPU. If it happens again and I get a minidump, I will post it.


My GPU is an Geforce GTS 450.

 

[smarteyeball]

So my question is, is it normal for my computer to be overheating with a GPU at 72? Or could there be something else underfoot that is causing my issue?

That temp is fine, so it's unlikely that heat is an issue. (The GPU core temp anyway)

I'd go with something underfoot - but without a crash dump, it's hard to offer a definitive alternative.

If it's only after an extended period of time, have you monitored your CPU's peak temps?

Have you updated your drivers during this period, or is it the same driverset?

 

[GeneO]

My computer fairly consistently BSODs while playing games for extended periods of time. Usually while playing a game my GPU peaks at about 70~72 C. I know this is up there, but I've seen lots of people not having to worry about overheating until they get to 90 or so.

So my question is, is it normal for my computer to be overheating with a GPU at 72? Or could there be something else underfoot that is causing my issue?

Unfortunately, the last few bluescreens I had were too quick to save any minidumps, and all my old ones are gone... I think I may have accidentally wiped them with CCleaner or something. I've been having this problem for a while and have finally (I think) narrowed it down to my GPU. If it happens again and I get a minidump, I will post it.


My GPU is an Geforce GTS 450.

That is kind of high for that GPU. I have the same and fully stressed I have the fan control set so it doesn't go over 60. I would advise using the MSI Afterburner program - it works for most any Nvidia and I use it for my Asus GTS 450. It allows you to create a fan response profile, which as I said I use to keep the GPU under 60 event overclocked under load

 

[Siesna]

Smarty, my CPU never gets over 45C or so. It never gets very hot. It's the same drivers, so next time I play anything, I'll just let it run until it bluescreens, and hopefully get a crash dump this time.

Geneo, thanks for the advice. I will check it out and try to keep my GPU around 60 and see if that helps. After I started getting these blue screens a while back, I even installed two more fans pointed at the card to hopefully cool it down, and it only helped by about 5 degrees or so, dropping me from 75 to around 70.

By the way, none of my hardware is overclocked. This is all stock.

 

[essenbe]

This will enable you to have mini dumps on BSODs. But, those GPU temps should not be high enough to cause BSODs.

 

[Siesna]

Thanks for the tip about MSI Afterburner. My GPU is running about 50C instead of 70C. Even though those temps are not hot enough to cause my problem, I like my system running cooler, regardless.

I got a BSOD again with my GPU only at 50C this time, so, definitely not an overheating issue. But once again, no dump file was generated. I don't know why. My system used to create small dumps in the minidump folder, but just now when I went to investigate, it had somehow gotten changed to a kernel dump. I changed it back to minidumps. However, no MEMORY.DMP was generated.. The BSOD is weird. The blue screen is only up for a half a second, then my system restarts, almost like it didn't have time to dump anything. When I restart, windows does not acknowledge the fact that it shut down unexpectedly. The event viewer only states that "The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly."

Trying to cause another BSOD now that I have told it to save small dumps.

Obviously this isn't a simple overheating problem. I only thought it was because my system is only restarting when I've been playing a game for a while.

 

[GeneO]

Well good then, you can probably rule out temperature. Is the BSOD error code, address and faulting module the same every time? Can you capture those an post them?

The are probably in your system log. Open the event viewer to the system log and filter on error, warning and critical messages.

What Nvidia driver version are you at.

 

[Siesna]

Event viewer is quite unhelpful. Not even acknowledging that it was a blue screen, when clearly I saw one when it restarted.

Whenever I've had a restart, the only relevant thing listed in the event manager is citing Event 41, being a Kernel-Power error. "The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly."

I'm just going to keep using my system like usual for now in hopes that I will get a blue screen that actually generates a helpful dump file. I've been having this issue for going on 6 months, but I gave up on figuring it out for a long while. I know I *was* getting dumps, but I apparently accidentally deleted them all. So here's hoping for a new one that will be useful.

My driver version is 290.36, by the way.

 

[panais]

Well....we need the minidumps.

In meantime,

290.36 is beta,install and use an older driver,for the moment dont use any beta driver.

Download and install Speccy-Free-Version take a snip of the summary and upload it here.

Download,install and run OCCT-video-card-test. Finish,find the OCCT test results folder in your documents,zip the folder and upload it here.
During occt stress\test If you see artifacts, or the temperature of your GPU exceeds 80C, stop the test.

Run http://www.sevenforums.com/tutorials/105647-ram-test-memtest86.html
Let it run until at least 9 passes are completed, or errors are found (whichever comes first). The longer you run it,the better.

 

[GeneO]

Or the latest beta, 295.51.

 

[Siesna]

...Well. I'm getting more and more confused.

Including the snip of Speccy, but I cannot get my computer to create a minidump, or bluescreen even now.

Every time I blue screened, either through OCCT, Furmark, or just playing a game, my computer restarts too quickly for it to save the dump. I figured my issue was that I needed to disable auto restart. So I did. And now my computer won't blue screen at all, but just makes the entire system go to crap. Windows Explorer assumingly crashes, since my start bar and desktop disappears, Chrome starts giving me this screen with every new tab I open, and any other program I have open (including OCCT) gives me various error messages basically boiling down to that it can't save anything on the hard drive.

I know I *used* to get minidumps. I just accidentally deleted them all with CCleaner. But now I can't get any new ones. It seems if I enable automatic restart, I will blue screen, but not save a minidump. If I disable it, after a while instead of blue screening, my computer just cacks up so much, I have to manual restart.

Some things I should have mentioned the first time:
My GPU and hard drive have both been replaced. They were under warranty still and I sent them in hoping that new hardware would fix the problem. Neither one did.
I have reinstalled windows.
I have ran both sticks of my RAM through memtest, about 14 passes each. I tested each one individually overnight. Both had zero errors.

I updated to the newest NVIDIA beta, but since it didn't seem to help, I'm going to go ahead and just install the older, not beta driver.

GAH.

 

[essenbe]

Take a look at this tutorial on how to enable dumps. maybe that will help the process.

I would also consider a PSU problem. Considering the fact that the BSODs occur under strenuous activity and the brand of CPU you have.

 

[Siesna]

That is actually a tutorial I have gone through, and followed it's steps exactly. Neither the kernel dump or small dump will generate anything.

Finally got a BSOD with automatic restart disabled, and this happened:

Computer sat this way for several minutes without advancing through the process of actually saving the dump... Finally had to manual restart.

I had not considered a PSU issue. I will see if it is still under warranty, and perhaps try getting a new one.

 

[essenbe]

Take a look at this thread from Microsoft Answers and see if there is any help there. You will be better able to determine if this fits your situation.

 

[panais]

If the GPU temperature 58C is from idle state then is very hot.

 

[Siesna]

GPU idles at around 30. It runs between 65~75 when under load.

I do not think my problem is the one described in this thread. It has nothing to do with my computer being in hibernation/sleep, but that is the same error code I got, so I went through the steps listed there anyway.

Here's the results of my check disk:



Checking file system on C:
The type of the file system is NTFS.

A disk check has been scheduled.
Windows will now check the disk.

CHKDSK is verifying files (stage 1 of 5)...
256768 file records processed. File verification completed.
211 large file records processed. 0 bad file records processed. 2 EA records processed. 60 reparse records processed. CHKDSK is verifying indexes (stage 2 of 5)...
318694 index entries processed. Index verification completed.
0 unindexed files scanned. 0 unindexed files recovered. CHKDSK is verifying security descriptors (stage 3 of 5)...
256768 file SDs/SIDs processed. Cleaning up 178 unused index entries from index $SII of file 0x9.
Cleaning up 178 unused index entries from index $SDH of file 0x9.
Cleaning up 178 unused security descriptors.
Security descriptor verification completed.
30964 data files processed. CHKDSK is verifying Usn Journal...
The USN Journal entry at offset 0x7196d600 and length 0x2e0073 crosses
the page boundary.
Repairing Usn Journal file record segment.
34470496 USN bytes processed. Usn Journal verification completed.
CHKDSK is verifying file data (stage 4 of 5)...
256752 files processed. File data verification completed.
CHKDSK is verifying free space (stage 5 of 5)...
178597016 free clusters processed. Free space verification is complete.
Windows has made corrections to the file system.

976760831 KB total disk space.
261889396 KB in 170612 files.
93960 KB in 30965 indexes.
0 KB in bad sectors.
389411 KB in use by the system.
65536 KB occupied by the log file.
714388064 KB available on disk.

4096 bytes in each allocation unit.
244190207 total allocation units on disk.
178597016 allocation units available on disk.

Internal Info:
00 eb 03 00 73 13 03 00 7b 9b 05 00 00 00 00 00 ....s...{.......
7c 2a 00 00 3c 00 00 00 00 00 00 00 00 00 00 00 |*..<...........
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................

Windows has finished checking your disk.
Please wait while your computer restarts.

Going to restate just in case:
The HDD has been replaced
Both sticks of RAM have been memtested for 14 passes each, tested separately.
The GPU has been replaced as well
Windows has been reinstalled.

I am going to try unplugging all unnecessary hardware, and using a different SATA cable for my hard drive to see if that may help. These are, however, all steps I have tried in the past but I will try again.

Is there any chance I could have a bad MOBO?

 

[GeneO]

Check your disk at a low level scan for errors. Use the seagate or western digital disk check software.

How much pagefile do you have setup? That error could indicate a disk issue where the pagefile is getting corrupted or insufficient pagefile.

The GTS 450 will not only use on-board, but also system RAM. This may cause swapping and hence be hitting your pagefile triggering this issue.

Also, since you had file system inconsistencies, I would run, from and elevated command window, sfc /verifyonly to check the consistency of the system files.

 

[Siesna]

My pagefile is 4 gb. Should I change this value?

Running SeaTools right now, and I will run the system file checker after that.

Also, I unplugged everything but fans, HD, RAM, and GPU from my mobo, and changed the cable/mobo port that my HD was plugged into. Still got a bluescreen when running OCCT, this time with error 0x0f4. Googling this error code led me to a thread of someone with a different Geforce GPU getting both 0x7a and 0xf4 like me. It was determined that their PSU was not sufficient to run the GPU. My PSU is a CoolMax CUL 650W. The specs are listed there.

This is the first rig I've ever built, and the PSU was one of the things I was confused the most about.

 

[Siesna]

Also, the HD passed the system file checker, and also passed every test in SeaTools.

 

[essenbe]

The PSU was mentioned earlier. If you have a friend who has an extra one or some way to test another PSU, I would give it a try before buying one without knowing if that is your problem. Most every other test you have done has passed. The problems occur when the system is under stress, which puts more demand on your GPU and PSU as well as other components. the problem is that you are not able to stress other components to find your problem.

Also, have you updated all of your motherboard drivers?