Closed Grzywax closed 3 weeks ago
This very much looks like issue #15 The Windows Terminal window has a "feature" where all execution will stop once you select anything within it, a simple click will be enough for this. And another click or a key press will then remove the selection and return to normal processing.
There doesn't seem to be a way to disable this using Batch or PowerShell, only the user itself can disable the "Quick Editing" feature (see the links in the issue mentioned above).
I saw that issue before posting. Point is that after starting script at 21h38 I did not touch my PC. I came back around 00h42, when I discovered that it's frozen / stuck on one core. Also last line before it froze says: Checking CPU usage: 27441.55%.
Could it be something else ?
ps. for sure there is a chance that I came up to PC around this time and I don't remember it. I was watching Foundation episode and in one shot - and I'm pretty sure it was 23h00 when I finished it. So I couldn't be at the pc at 22h16... but like I said, it's Friday evening, after long week at work... maybe my memory is playing trick on me ;]
At least it has all the indications of it. No log output or activity for CoreCycler whatsoever and then after being re-enabled it parses all the things that have happened in the meantime.
The CPU usage is indeed weird, and it may indeed point some other problem - but I have no idea what it could be that causes this spike and then the script to freeze and then re-awaken on user interaction besides the mentioned Quick Edit feature.
That being said, there are reports that Windows just "does" this from time to time, even if Quick Edit is disabled. So it may just have been one of these coincidences with no apparent reason. 🤷
So yeah... basically all those issues slowly go away the more my CO gets optimized. It seems that other cores ware causing the script to fail. Funny thing is that now I can run core-cycler for days and not throw any error on any of the cores, but every few days I would get random reboot/blue screen. So some cores/core are/is still set too high in CO but with core-cycler is now not enough to detect which one. I guess it would have to be super light fast task to would boost clocks to their max and give bad result... I hate Ryzen 7000. It tempts you with CO potential but then to get it perfectly tuned is like fighting windmills.
Closing this, as it's impossible for me to reproduce. If you see this happening again, let me know.
Hi, first of all great job! I started to lay with this script and it's great. The fact that even Ryzen Master does not test in similar way to set PBO is beyond me. It often baffles me why obvious solutions are created by individuals like you.
Anyway I have this weird behavior where when I leave scrip unattended sometimes it freezes. And I have feeling that it only continues when I wake up my PC and click on CMD window - then it "wakes up". Also when it wake's up it spits out log for Pime95 for those 2.5h that was missing. It's like it got stuck on one core for some reason.
Here's fragment of the log -> jump from 22:16:06 to 00:42:30 (below I have pasted output as well + Prime95 log attached):
..... cut....
-> Output: