Is MSI Afterburner's Hardware Monitor accurate?

I know this sounds like a really stupid question to ask, but hear me out.

See I used to use nVidia's System Tools to OC. But for a good while now it had a habit of crashing at pretty much any given time, and quite constant. And some games crashed almost instantly, like Skyrim/Fallout. It's so bad in Skyrim that I loaded in once, heard someone next to me say 1 word, then it crashed.

Some would let me play a while, but had a high rate of crashing like COD 4 (not W@W/MW2/BO/MW3), CS:S, TF2 and HL2 (and all other HL2-esque games like EP1 & DM). It has nothing to do with temperature, as I can crash in HL2 at 65 degrees and play all day in Borderlands at 95 degrees (yes, borderlands loves making my GPU hot, it doesn't get exactly cooler than that no matter what I do).

I fixed this issue by simply not OCing my GPU anymore. Never crashed after that. But now I was talking to someone and I heard that if temperature was not an issue in the crash, then it's probably the memory having issues with a crappy software overclock. So I try MSI Afterburner and the same clocks (623/970/1550) seem tons more stable. I played skyrim for like 2 hours but then it crashed exiting a house. When I checked MSI afterburner, it said my GPU clock was running at 675, and my shader clock was running at 1674. The memory, however, was in lieu of my set options, running at 972. Now if this is true, (because it does not match up with what GPU-Z says, which is 623/972/1550), then my card is operating outside of safe parameters, as the highest stable clocks I've ever used are 650/1020/1650. Even a 660 core would almost instantly crash any game.

Has anyone heard of anything similar to this?

 

Really? Nobody knows? >_<

 

try the MSI subforums, I'd imagine they are a pool of people who are more likely to use it. Alot of them dont venture outside of there.

 

You might be right, but I came here because I'm using a mobile GPU and I'm SURE a lot of these people use MSI AB to OC their cards. I might just post it in the Sager/clevo forums, almost everybody who buys a clevo OCs their GPU at some point.

 

This sounds like a stability issue, not temperature. Have you run any real stability tests on it?

 

I used to run them way back when, I would usually pass stability tests like furmark/prime95 etc. But the problem is that it seems that no matter what I OC my card to, it will simply crash. If my card simply has a problem with software OCs + Windows 7 (as I didn't have these issues before I changed from Vista/system tools 6.03 last year) then I'll simply leave it be until I can somehow get a new machine with a newfangled 680M card (that I probably won't even need to OC anyway).

But if it forces my card to run at higher, unstable clocks for some reason (which would explain why an OC of a simple 610 core causes crashing, as according to MSI afterburner it would be operating above 650 core speeds) then that might be something I can fix, or simply roll with and make tiny core/shader OCs while pushing the memory a bit higher. I just want to find the root of the problem, or if there is simply none that I can rectify and the issue is MSI Afterburner giving me faulty readings, then I'll not bother with it at all. I can't stand not knowing, however.