I’ve a really annoying bug on my iMac 2019 (27″, Retina 5K, i9, 8cores, Monterey, Radeon Professional Vega 8 Gb, 64GB Ram). It is possible attributable to a {hardware} drawback but additionally triggers a software program bug I might attempt to mitigate.
The symptom
Generally, and at the moment almost on a regular basis, my Mac turns into unresponsive, the fan will get caught to 3600 rpm, and might keep so for hours.
The analysis
This may be traced to the Mac making an attempt to chill itself by launching a kernel_task course of, which is a identified technique when CPUs are heating, devised to exclude different offending processes.
I’ve put in a some third-party utility and might reproduce the issue:
Fairly at random (even in the midst of the night time when the pc is asleep), a temperature sensor (named AC/DC T2 Sec HeatSink) would go in seconds from a traditional room temperature to 70, 80 and even 125°. All different temp sensors present regular values (<40°) and no spikes.
Then kernel_task kicks in, often round 1000 %, making the Mac unresponsive then the fan goes to most velocity (3600 rpm).
After some time, typically in few seconds, the sensor often comes again to regular by itself, and it additionally does so if I knock bodily on the place the place it’s situated (close to the plug).
I’ve studied extensively console logs and couldn’t discover any particular course of that will set off this behaviour. I’am very inclined to assume it is a {hardware} bug, and that is additionally Apple opinion, since I’ve already introduced my Mac for restore because of this twice, they modified the logic board, and it disappeared for a couple of monthes.
Treatments and workaround
I certainly ought to return to Apple now a 3rd time, however I am questioning if I may delay it, particularly since they charged me the final time, for what I imagine is a design fault. Final time they weren’t of nice assist in understanding the issue, telling me I had an excessive amount of third-party reminiscence and to return to 8GB Apple.
Additionally I do not actually need to improve my OS, since it will break some third celebration software program once more. For the report the issue was right here on Catalina and Large Sur and upgrading the OS introduced nothing so far as this bug is anxious (besides from third-party obsolescence).
At the moment, each time this occurs, I knock on the Mac to convey the sensor again to regular worth if wanted, then I droop exercise (utilizing Apple menu), watch for the time for it to occur, between a couple of seconds and a minute, then I get up the Mac and all is again to regular.
Clearly there may be additionally a software program bug within the OS, since there isn’t any course of to chop kernel_task and produce again the fan to regular values when all temperatures are reported inside regular operation vary.
I do not know the basis reason for the {hardware} bug. I present some clues for the reader curiosity:
-
Have I broken the sensor by unplugging reside the Mac (this will occur unwillingly if you transfer the display screen, the wire plug being caught on the wall behind) ?
-
Is there some gathered mud that produces a short-circuit and strikes after I knock ? I can not clear it, the iMac cannot be opened
-
Is there too many USB units plugged and does this increase the power wanted, triggering the sensor (there’s a loosely correlation with the activation of the USB DAC related, however can also be occur when it isn’t right here) ? I plan to purchase a self-powered usb hub, simply in case.
Which brings me to the target of this submit and my query :
May I cease the un-needed cooling course of with out suspending exercise ?
There are some utilities that cease the fan however hey do not calm kernel_task, so that is of no use.
Making an attempt to kill the kernel_task course of appears a nasty thought, as it’s important to the system.
There are clearly under-the-hood actions that treatment the issue when exercise suspension is triggered, so I ponder if I may attain the identical outcome with out loosing the minutes I at the moment unfastened due to the exercise suspension.
Thanks on your inputs.