• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Irregular frequency with Windows 10 (bus speed?)

Joined
Feb 16, 2012
Messages
415 (0.09/day)
Location
Sweden
Hey guys. I have noticed this for a while now but since i upgraded to Windows 10 at its release my overclocked frequencies has not been steady. This is exactly the same settings I had in Windows 8.1 which didn't have this "problem". I have also measured the same results with both HWINFO64 and Aida 5.50.3600.

Is this just a measuring problem in Windows 10 or is there something else to it? Something I have missed maybe? (Since voltage reading is not working in CPU-Z)

First and foremost, here is my current settings.

4960X
4.5Ghz
45 All core Turbo
1.25V
100 Mhz BCLK
Every C-state Off
Speedstep Off
CPU Spread Spectrum Disabled
Maximum performance in Windows power settings.


For example, when its idling in Windows i get a result pretty much like this. The frequency is bouncing around according to the bus speed which changes between 96.7 MHz all the way to 99 MHz.

k8bDYTN.png


And when I run the built-in benchmark in CPU-Z, single core results looks like this:

64kFIRN.png


Looks almost exactly like the idle one, except that the Bus Speed is 96.8 MHz. And finally, when all cores are in use it jumps up a bit, but it still does not hit that BCLK I have put in place.

274TCkT.png
 
I get the same thing with my 6700k.
 
Hyper-V is known to cause this behaviour.
 
Hyper-V is known to cause this behaviour.
Wow, you are completely right. I uninstalled Hyper-V and it jumped back up to where it should. Since I have a server for this purpose now I don't need the feature.

I would never have thought of that. Thank you very much for the help.
Not that this really was a performance problem more of a confusing OCD-thing going on.

Does Hyper-V really affect the frequency or is it a sensor issue?

E16K92T.png
 
Last edited:
Well, do you have any machine running in background? That would be the first place I would look.
Unless you have explicitly shut it down, not just closed, it will always run, even after PC restart.
 
Does Hyper-V really affect the frequency or is it a sensor issue?
I've read a couple posts on various forums that indicate that more than one device accessing the timer will cause strange results with the readout, not with actual clocks. Although I'm not going to lie, this being the issue blows my mind and I'm not sure if the two are related. Extra props for @ricoh on this one.
 
Well, do you have any machine running in background? That would be the first place I would look.
Unless you have explicitly shut it down, not just closed, it will always run, even after PC restart.
Im really sure I didn't have it a machine running. I think i removed them completely when i transferred the machines over to the server.

I've read a couple posts on various forums that indicate that more than one device accessing the timer will cause strange results with the readout, not with actual clocks. Although I'm not going to lie, this being the issue blows my mind and I'm not sure if the two are related. Extra props for @ricoh on this one.
It is very interesting, if it affects it even without any machine running.
 
I get the same on my 4690k, jumps up to 102 bus....I believe it is just a conflict when running two bits of software monitoring the same thing but gonna disable hyper v and will share if it fixes it.
 
I will have to check my own PC for this, since I use Hyper-V client.
Let's see what happens with no VM's running.
 
Back
Top