kernel_task high CPU - faulty thunderbolt port?

There's a slight chance this might help:

Resetting the System Management Controller (SMC)

So this has been happening to me too... it's been infuriating as it's neither consistent, nor easy to resolve once it starts happening. Restarts and SMC resets SOMETIMES do not resolve the problem.

Thank you, rodogu, for your insight to this. Here's some further information. This happens to me BOTH at work with a ThunderBolt connected LED Cinema display (the model before the Thunderbolt Display), AND at home on an older 23" Apple Cinema Display connected, by adaptor, to the Thunderbolt port.

My MacBook Pro is actually an 8,3 (2.3GHz i7) running 10.8.3.

After reading your post I unplugged the display from the Thunderbolt port (leaving the displays USB connection plugged in) and presto!... CPU usage on process 0 (kernel_task) went from 609% to basically nothing.

APPLE FIX THIS!!!!!!! Sheesh..........

i've been experiencing the same thing on my early 2011 15" macbook pro for quite a while now. the cpu usage by kernel_task will go up to 260% without any other actitive task on the machine while connected to an external monitor via thunderbolt/DVI connector. this is really annoying. i have tried SMC reset a couple of times and it didn't solve my problem. APPLE, i would at least want to know what exactly is wrong with my macbook.

Have you tried unplugging and replugging in your display? When I do it (temporarily) resolves the issue.

it does, but only for a short time like a minute or two. the next time i scroll around in the browser and the cpus are on high load again :-(

weird... the differences between our situations are it seems like it doesn't necessarily tap as much of your CPU as it does mine (although I don't know what CPU you have... maybe it does)... mine is 75-80% which slows me down so much that sometimes easy things like text entry can have a 15-30 SECOND delay before letters start appearing one at a time.

I rarely have the problem return so quickly... in fact it hasn't happened yet today for me (although I know it will).

Okay, I found out what was the issue. The macbook was overheating, becasue of a macbook cover that I had on it.

Aparently when tempature rises that is not releated to CPU, kernel_tasks generates load on the CPU. Load on the CPU spins the fans that cool the macbook down. In my case the cover forbidden to cool it quickly enough so the slowness of the OS started to be noticable.

Using external display increases temputere of the GPU inside the macbook, however CPU load is low as the external display doesn't make much diffrence for the CPU. Since it's low, fans don't start, so karnel_tasks generates random load on the CPU to trigger the fans and cool the system down.

It's not Thunderbolt issue or Thunderbolad Display problems. You should all check why your macbook may have problems to cool itself down.

Really Apple? Can you please supply a patch that stops kernel_task from doing so and instead tune the fans manually. There are apps like Fan-Controle so it is possible to spin up the fans without wasting my CPU.

A simple question: if a house burns do you call the firedepartment or start a fire in your house, so they might see it themselves.

Sounds like a plausible theory, but how strong is the evidence you have to support it? It's rather Rube Goldberg.

I see no reason why the system wouldn't simply start the fans directly when any of the temp sensors indicate they're needed, rather than tie them to the CPU temperature only.

I've been having this problem with a Retina MBP attached to two external Apple Thunderbolt monitors (Monitor A -> Monitor B -> MBP). kernel_task pegs the CPU at 600%+ and stays there, making the system mostly unusable. Unplugging the monitors allows the system to recover and kernel_task CPU usage drops to around 10%.

I shut my system down then used a compressed air can to blow out the Thunderbolt ports on the MBP and the end of the Thunderbolt cable coming from the monitor. I reset the PRAM and SMC then made sure the Thunderbolt cable was seated properly in the port and rebooted.

For the moment this seems to have fixed the problem.

I don't think this is related to thermal problems. I don't have a cover on my MBP and run it with the display open. I didn't have anything intensive enough going on to drive the heat up except for kernel_task going crazy.

Just wanted to share my equally depressing experience.

I'm using a Late 2011 MacBook Pro 17" with 16 GB of RAM, a 256 GB SSD Drive, and running Yosemite. The MacBook runs fine on its own, but when I connect the Thunderbolt display to my 2011 23" iMac running Mountain Lion and use Target Display Mode, things work fine for a few minutes, but once I start using any of my music programs (Logic Pro X or Ableton Live 9), or even Chrome, the MacBook fans will blast and Kernel_Task will take up a whopping 400-500% of my CPU! It's absolutely insane and devastating to my workflow. It's so important that I am able to use an external display when I'm in my studio.

I have yet to find a solution to this, but I will keep looking...

I'm running a Mid 2012 MacBook Pro Retina...

Have been having this problem with the monitors at home where I have 2 Thunderbolt Displays.

It appears, at the moment that the problem is the MagSafe 2 power adapter.

I tried it on BOTH displays so far and as soon as it's plugged in and giving power, to the laptop through the adapter it takes about 3 minutes and the Kernel_Task process spikes. So far I have been running it for about 20 minutes on the laptops power adapter and it seems to be fine. Haven't had it spike the CPU at all.

One thing I did notice, is that I kept having to bump the power adapter (Thunderbolt with MagSage 2 Adapter) to get it to power the laptop. That's out I found it. But when I took the power adapter off, it was VERY warm.

Try it, it might be the problem.

ok, one other thing you should try is to go get the latest Combo updater for the version of the OS you are running.

Here is the link to the 10.11.4 update.

Download OS X El Capitan 10.11.4 Combo Update

Applying this also seemed to have cleaned up a lot of issues on my machine as well. I am still testing out both solutions. I did get it to start spiking again last night after 4 hours on the normal adapter, But it hasn't happened since I applied the combo update... Yet.