Then do some browsing, opening/closing assorted tabs, Notepad, Word, Chrome, etc; after 5 minutes, check HWMonitor tab and see what the max clock speeds were achieved...
At least one of the cores should, at some point, boost to max during single threaded scenarios, as described above...
If you then try CPU-Z/bench/stress CPU you should see lower clock speeds (and a decent core/package temp) as all cores are run at ~90% load....; when the test is stopped, clock speeds/temps should both rapidly drop back down into the upper 30's-mid-40's when at near 800 MHz or so..(true idle condition, when no WIndows tasks are repeatedly stepping in, causing clock speeds to ramp up for a quarter second or two, which is normal...)
I just put this system together and was confounded by my 3900x temps. Idle temp was high and frankly "cpu" temp it didn't really seem to respond to load in any sort of useful way. (making a good fan profile was laughable)
I started experimentation and found that the "cpu" temp had really nothing to do with the core loading. It was mostly memory and IO related:
Idle at 59-63*C, Mem at 1.38V, 3800MT (XMP)
Idle at 51-55*C, Mem at 1.25V, 3800MT (XMP)
Idle at 41-43*C, Mem at 1.20V, 2400MT
Idle at 39-41*C, Mem at 1.10V, 2400MT
Loaded 78-81C and slowly climbing, Mem at 3800 MT (XMP) 1.38V
Loaded 71-80*C and slowly climbing, Mem at 3800 MT (XMP) 1.25V
Loaded 68-71*C and slowly climbing, Mem at 2400 MT 1.2V
Loaded 65-69*C and slowly climbing, Mem at 2400MT 1.1V
My past experience says pushing a silicon chip connected to a metal heatsink results in a increase in temp over time. With this 3900x, the reported CPU temp increase and decreases instantaneously with load. Really! instant on both the up and down slope... crazy.
IMO this could only happen if the CPU temp was either A not connected to a heatsink, or b some sort of fake algorithm.
When you have three chips making up a "CPU" what do you report as the temp? ... AMD has clearly messed this up somehow.