Could you please attach a screenshot of the HWiNFO sensors window to check what is displayed there?
Also, is there any other tool displaying 1600 MHz for the 3rd core? I'd also need the HWiNFO Debug File from this situation to check it in more detail.
Thanks. However the debug file shows all cores at 16x ratio (1600 MHz) during sensor monitoring. Have you created that dump during the situation you described originally ?
Thanks. According to this dump all CPU cores report 41x-42x ratio, so they really seem to be running at the clock reported by HWiNFO.
This is indeed strange because Core3 doesn't seem to be utilized according to system utilization status.
When you run TMonitor which reported Core3 at 1600 MHz, have you did that in parallel with HWiNFO ?
Also, can you please try to disable monitoring of some sensors (like SMART for example) in the HWiNFO sensors screen (right-click on sensor and Disable Monitoring) if that maybe changes the situation ?
I've disable SMART, no changes. I ran TMonitor alone.
I must say I'm getting wrong values in CPU-Z and AIDA64 as well. TMonitor is the only giving correct values. This is very strange, I don't have a clue of what's happening!
Thanks for your attention!
When you run both HWiNFO and TMonitor together, does TMonitor still display Core3 at 1600 MHz ?
Thanks. This is indeed a bit strange, so I'll check this with the author of TMonitor and will let you know...
Well, so it seems that if a core is in idle state, then on certain CPUs (depending on settings), just a slight workload on that core can bring it to a raised performance state. That means if a tool like HWiNFO just tries to query the actual parameters (including current multiplier), it can cause the core to go out of idle state for a very short time.
There's a level of sensitivity to which the CPU can be set and I assume your system is set to a very sensitive level. If you check your BIOS, there might be a setting like "Energy performance BIAS", or something similar. Changing this might reduce the sensitivity and other tools to report correct values...
I'm afraid, I don't have a solution for this currently.
Changing to the way how TMonitor calculates the clock would bring other disadvantages and being unable to report other parameters.
That's ok, Martin, I understand!