5ghz A8 5600k?

P

pedanggila

Member
Spec:
A8 5600k (Extreme profile LLC, Turbo Core off, 44x, 1.453v)
Gigabyte F2A85X UP4
KVR16N11/8 (1866mhz 10-11-10-29, 1.55v)



Seems like it's quite off the chart or it did actually reached 5ghz, my a8 5600k. Temperature is normal (for ambient 30°C tropical climate) and voltage only maxed at 1.5v. Is this possible? But if it really did maybe it's just there for nano secs.

[attachment=765]
 

Attachments

  • bug.jpg
    bug.jpg
    110 KB · Views: 17
The final settings was done thru bios, win7 ultimate and if the FBC is by default ON then the only thing I've changed in Config was enabling the debug mode. So yeah here's the dbg file. It was compiled after I got the 5Ghz mark so maybe I missed something for a proper monitoring setup.

Thank you for the fast reply :)
 

Attachments

According to the data provided, the maximum ratio the CPU reached is 45x and clkin (bus) = 100 MHz.
So I have to admit, that this is a total mystery to me.
Could you please use the latest Beta build (v4.23-1990) and perform more test runs and watch the maximum value - what's the exact maximum clock reported and if it's the same on all cores ?
 
Does it always show the exactly same max clock value (5159 MHz) and on all cores ?
 
I can't tell when or how long it stayed at 5ghz since it's running at the background and it must've triggered during gaming, so yeah no direct observation. It was after I exit the game and only then I noticed HW gave me a weird figure.

Another 5ghz reading came out but this time along with 2GHz NB frequency with very high NB voltage reading of 1.238v. 1.550v vcore is innacurate though as it should be max 1.5v.

[attachment=783]

I wonder all these spikes are related to Extreme LLC profile set in bios but the number doesn't add up, like you said 45X multiplier with 100Mhz bus clock should not equal to 5Ghz.
 

Attachments

  • sp.jpg
    sp.jpg
    103.9 KB · Views: 4
Does the Debug File you attached before really capture the run when the spike occurred ?
I believe the game does somehow interact with the (bus) clock measurement method in HWiNFO. I also think other tools (like CPU-Z) might be affected as well, but there's no way how to check the history of values recorded (max) there.
I don't have a solution for this right now, but I think I'll implement a new setting in HWiNFO which when used will measure the bus clock only once during runtime, so later when starting an application (which might interfer with it), this won't happen.
 
The next release will add a setting called "Periodic polling", which when disabled will help to mitigate/fix this problem.
 
Currently testing 4.24v. With Periodic polling ON, that strange number did came out but when it's disabled, though with just 20 minutes of test time or so, everything's normal. So far so good I think. I'll report if there's any anomaly.
 
You must log in or register to reply here.
Back
Top