Stuttering in Windows 7 x64

[Khendraja'aro]

Okay, here's the problem: Whenever I play a game, it runs smoothly for an arbitrary amount of time (~3 minutes) and then it begins to stutter. It looks as if the framerate cuts down to about ~5FPS and the sound stutters as well. This stuttering then continues for about 30, after which everything returns back to normal. And the cycle begins again.

This is independent of the actual hardware demands of the game. I've had it happen in Torchlight while I was in town. I've had it happen in Crysis 2 while running through a tunnel where absolutely nothing was happening. I've had it happen in Warhammer 40K Space Marine while wading through Orcs or watching a cutscene. The same cutscene will play fine if I restart the game - the stutter will set in at another unrelated point.

I've already looked at the usual suspects (drivers, temperature, 3rd party processes, Aero, ...), so it's nothing that obvious, okay?

I've also reinstalled Windows completely, it's still happening on a default system.

Only thing that I'm seeing is that during those stutter times, both cores will be at 100%, like this:



However, as you can also see, the processes don't reflect that CPU consumption - essentially, they all stay at the level they were before the spike.

Info on my system:
AMD X2 6000+ @3.1 GHz (not overclocked)
4GiB RAM
AMD Radeon HD 5570
ASUS M2N68-AM PLUS

Any good ideas, anyone?

[Ominous Gamer]

Do you have AMD's Dual Core Optimizer installed (even as a service)? Any throttling programs (even in the BIOS) like Cool and Quiet?

[Khendraja'aro]

Dual Core Optimizer is not needed for Windows 7 (and the latest version is from 2008 anyway).

Will have a look at Cool'n'Quiet, though I used CPU-Z and did not see a reduction in frequencies anywhere.

Also: Turning vsync off did not help and neither was my dual-monitor setup the culprit.

[Ominous Gamer]

It may not be needed, but its installed by default via some drivers. I know it ends up as a service on my Windows 7 Ultimate Sig Edition after I let Windows Update run through all its updates, driver downloads, and patches.

[Khendraja'aro]

It's not Cool'n'Quiet either.

[Khendraja'aro]

And it's not the Realtek onboard sound either. Damn.

There's a whole lot of people with this kind of problem out there - I'm seeing few solutions or circumstances which fit to my problem.

Weird thing is that I did not change a thing about my PC in the last 6 months.

[Khendraja'aro]

Think i nailed the problem - seems as if my particular processor's temperature reader may be faulty, thus giving wrong readings.

Cleaned the thermal sink and the processor of old thermal paste and applied new one. No stutter since.

[Nessus]

Cleaned the thermal sink and the processor of old thermal paste and applied new one. No stutter since.

I once wrote a program for an experiment of one of my friends. The idea
was to simulate, in a computer, families of neutral beings; they would be
homeostats, cognizant of their "environment" but possessing, initially, no
"emotional" or "ethical" qualities. These beings multiplied -- only in the
machine, of course, therefore in a way that a layman would call
"arithmetically" -- and after a few dozen "generations" there continually
appeared, over and over again, in each of the "specimens," a characteristic
that made no sense at all to us, a sort of equivalent of "aggression." After
many painstaking but fruitless checking calculations, my friend, at his wit's
end -- really grasping at straws -- began examining the most trivial
circumstances of the experiment; and then it turned out that a certain relay
had reacted to the changes of humidity in the air, and thus those changes had
become the hidden producer of the deviation.
I cannot help thinking of that experiment as I write, for is it not
possible that social evolution lifted us from the Animal Kingdom in an
exponential curve -- when we were fundamentally unprepared for the ascent? The
socialization reaction began when the human atoms had barely given evidence of
their first cohesiveness. Those atoms were a material strictly biological, a
material made and prepared to satisfy typically biological criteria, but that
sudden movement, that upward shove, seized us and carried us off into the
space of civilization. How could such a start not have bound onto that
biological material accidental convergences, much as a probe that, lowered to
the ocean floor, scoops up from it, along with the desired object, debris and
chance pieces of junk? I recall the damp relay in the sophisticated computer.
And the process that engendered us -- why, pray, must it have been in every
respect perfect? Yet neither we nor our philosophers dare consider the idea
that the finality and singularity of the existence of our species do not at
all imply a perfection under whose aegis the species originated -- just as
such perfection is not present at the cradle of any individual.

[Lebanese Dragon]

You should post your result on some of the websites your browsed, if they already don't have that answer... Kinda lame though how much they'll have to dig into their computer to see if they're sharing the same problem. I think there is free software that can give you a second opinion on the temp though, if your reader is off.

[Khendraja'aro]

You should post your result on some of the websites your browsed, if they already don't have that answer... Kinda lame though how much they'll have to dig into their computer to see if they're sharing the same problem. I think there is free software that can give you a second opinion on the temp though, if your reader is off.

Yes, software is perfectly able to solve a hardware problem...

[CitizenCain]

Any good ideas, anyone?

[...]

Yes, software is perfectly able to solve a hardware problem...

If you're gonna be like that, I don't know why anyone would bother... but since I'm such a nice guy, just replace your computer next time, given the complete inability of software to influence hardware problems.

[Lebanese Dragon]

Yes, software is perfectly able to solve a hardware problem...

how old is this kid? Did you not understand my post? If you downloaded a new temp reader (software), that actually read your temp correctly, you'd know that your CPU was getting overheated (so you could solve the hardware problem). If it comes up the same as your current reader, it doesn't tell you anything at all, but if the new reader comes up high it tells you somethings wrong with your temp.

I am curious though as to why you felt the need for a snobby response?

[Khendraja'aro]

how old is this kid? Did you not understand my post? If you downloaded a new temp reader (software), that actually read your temp correctly, you'd know that your CPU was getting overheated (so you could solve the hardware problem). If it comes up the same as your current reader, it doesn't tell you anything at all, but if the new reader comes up high it tells you somethings wrong with your temp.

I am curious though as to why you felt the need for a snobby response?

a) First of all, I already solved the problem
b) Now, that I solved the problem, I don't have a baseline to compare to. Making any "new" software equally unreliable. If the hardware sensor is faulty, it's a bit of a beast to come up with a solution without having anything you can compare your readings to. After all, the relation between temperature and resistance need not be a linear one.

[Lebanese Dragon]

You're not reading my posts correctly, glad your problems is fixed. Now you just need to fix your lack of etiquette.