Originally Posted by BroadPwns
Now do some proper testing that shows an actual performance difference.
Performance wise in CB20 with HPET on meant only a 1.6% hit (measured) In games It was within the margin of error. So not change there.
My task manager is way more jumpy with HPET on, I don't know if that is a good or a bad thing. With HPET off the cores are much more relaxed.
I'm running a fixed OC and I'm very aware about the background apps/services running, boot after boot I can repeat the very exact levels of usage 1min after boot, it is literally as in the images.
But, with HPET on, if I have Chrome opened with youtube I see more usage, a really noticeable difference. You have to run those commands and check for yourself.
I ran/run real time audio applications for the last 25 years and for me low DPC latency is paramount.
My test file showed I could run 198 instances of a plugin with HPET off (BTW, a 9900K could only about 165), and it was badly dropping out with HPET on. So for me indeed it makes a big difference.
To me this is the one that made me say "nope, I'm staying with HPET off" because the DPC latency checker gave me a consistent warning just after 10s of monitoring DPC, it was jumping up and down all over the place with HPET on. That with Chrome closed.
With HPET off it is hovering around just 40ns most of the time. I can leave it monitoring and do whatever in Chrome, forget about it for 5min and see no warnings.
It seems like is gonna stay off in my system. I don't care what it does, my 1800X had it off for 2 years and it was indeed beneficial there as well.
If you wanna try do so from windows, don't disable it in the BIOS.
Here HPET off resulted in no draft of time, I compared the system clock against my phone's clock for 20min and there was no draft. (just in case I didn't mentioned before)
BTW: No HPET off doesn't solve the CPU boosting in Zen2.