Question About Power Reporting Changes?
Question About Power Reporting Changes?
System details are as follows: 5600x display, RTX 2060, 32gb DDR4 RAM at 3200mhz, 2tb Samsung 980 Pro, M.IIO A550BN PSU, and MORTAR B550M Wi-Fi adapter.
Image reference: https://imgur.com/a/AfP64NR
View: https://imgur.com/a/AfP64NR
The first picture displays the system idle, and the second shows performance under load. I am worried about the Power Reporting Deviation reaching 113% during high usage.
This detailed analysis covers the reporting deviation metric in HWInfo.
As mentioned earlier:
It's likely that the reference value isn't correctly calibrated by the MoBo manufacturer (MSI).
If it reaches 100% or more, that's acceptable.
However, if it falls below 100% during near-full CPU load—such as at 75% or 50%—then there are some issues from the MoBo side.
Additionally, your value can reach up to 360.5%, which means the calibration of the reference value is significantly off.
I know the high idle readings don't make sense, but could the motherboard be signaling the processor it's using more power than it should because of performance problems?
So is everything okay on the board? I recognize that elevated idle values are irrelevant, but will the motherboard inform the processor it's drawing more power than necessary leading to slowdowns?
113% offset simply indicates the CPU believes it is using more power than it actually does, which can cause it to restrict the CPU to higher clock speeds. Your CPU runs at a base frequency of 3.7 GHz with boost capabilities up to 4.6 GHz. When you test its performance, does it reach 4.6 GHz? If yes, the offset shouldn't hinder its speed. However, if the boost climbs only to around 4.5 GHz or below (not exceeding 4.6 GHz), this may restrict performance because the motherboard signals the CPU it is using more power than it should, prompting the CPU to cap itself. Boost clocks above 4.6 GHz are flexible, so anything over 3.7 GHz base is normal.
Then, base and boost clocks are normal and nothing to worry about.
After enabling pbo on cinebench r23, the value remains approximately 4.3 to 4.35, which seems correct.