Yes I used the original hack code in the test.Thanks for reporting back your experience with it, good choice and useful to know. Just curious, you used the original hack code in testing, not the ActionTimer version, correct?
Not sure if a way to generate a load that will guarantee a certain CPU usage output would have been required to notice potential mismatch - I believe that placing the hack code results side by side with the standard 24 measures results would have been enough for an easy comparison. If the results were different, then the inconsistency in updating would have been clear.
Uhhh... NGL it did not occur to me to have then both loaded. You're right, that would have been a much smarter way to spot a mismatch.
Here I was staring at the meter and trying to see if task manager's little line graph changed. Was not a fun time.
But after the epiphany of loading 2 copy's, it seems every once in a while, one of the cores will disagree with the standard method and stay on its old value. But I do wonder if it isn't due to the timing of the update. Say the hack version updates ever so slightly slower or faster than the standard method and causes the discrepancy.
I'm running a X3D chip and its hard to load a core to a consistent say 40%. I did try a similar site but with Windows 11 having a mind of its own the per core usage isn't consistent. With modern OS and chip design I'm not even sure it's possible to load a core with such tolerance.A simple way to generate a load that is fairly stable is here: https://cpux.net/cpu-stress-test-online. Tweak the Threads and Power settings until your cores are running at the level you want. (If applicable, be sure to ignore the giant crapware ad at the top of the page with the Download button. This site might not function if you use an ad blocker, unfortunately.)
Statistics: Posted by Inheritance — Yesterday, 6:09 am