Kernel_task very high CPU usage

kernel_task running wild can be caused by some other process not functioning correctly. It can also happen if the cpu temperature is at its peak. The OS will throttle the cpu by taking up cpu cycles with kernel_task.

You can run Etrecheck to see what you have installed that might be causing issues. http://www.etrecheck.com/

Etresoft has instructions for how to post the results to this forum in the application help. Just Share Report to Clipboard (Copy Report) and paste the results in a reply here.

I have the same problem. I share the results of Etrecheck on My Mac

EtreCheck version: 1.0b11 (1A012)

Report generated: 2018-08-20 19:55:22

Download EtreCheck from https://etrecheck.com

Runtime: 1:33

Performance: Excellent

Sandbox: Enabled

Advanced Checks: Disabled

Adware Check: Disabled

Problem: Computer is too slow

Major Issues: None

Minor Issues:

These issues do not need immediate attention but they may indicate future problems.

Unsigned files- There are unsigned software files installed. They appear to be legitimate but should be reviewed.

32-bit Apps- This machine has 32-bits apps that may have problems in the future.

Hardware Information:

MacBook Pro (Retina, 15-inch, Mid 2015)

MacBook Pro Model: MacBookPro11,5

1 2.5 GHz Intel Core i7 (i7-4870HQ) CPU: 4-core

16 GB RAM - Not upgradeable

BANK 0/DIMM0 - 8 GB DDR3 1600 ok

BANK 1/DIMM0 - 8 GB DDR3 1600 ok

Battery: Health = Normal - Cycle count = 288

Video Information:

AMD Radeon R9 M370X - VRAM: 2048 MB

Color LCD 3840 x 2400

LG FULL HD 1920 x 1080

Intel Iris Pro - VRAM: 1536 MB

Drives:

disk0 - APPLE SSD SM0512G 500.28 GB (Solid State - TRIM: Yes)

Internal PCI 8.0 GT/s x4

disk0s1 - EFI (MS-DOS FAT32) [EFI] 210 MB

disk0s2 [APFS Container] 500.07 GB

disk1 [APFS Virtual drive] 500.07 GB (Shared by 4 volumes)

disk1s1 - H*******a (APFS) (Shared - 172.15 GB used)

disk1s2 - Preboot (APFS) [APFS Preboot] (Shared)

disk1s3 - Recovery (APFS) [Recovery] (Shared)

disk1s4 - VM (APFS) [APFS VM] (Shared - 1.07 GB used)

Mounted Volumes:

disk1s1 - H*******a 500.07 GB (325.58 GB free)

APFS

Mount point: /

disk1s4 - VM [APFS VM] (Shared - 1.07 GB used)

APFS

Mount point: /private/var/vm

Network:

Interface en5: Thunderbolt Ethernet

Interface en0: Wi-Fi

802.11 a/b/g/n/ac

Proxy Auto Discovery

Interface en4: iPhone

Interface bridge0: Thunderbolt Bridge

Interface en3: Bluetooth PAN

System Software:

macOS Mojave 10.14 (18A371a)

Time since boot: Less than an hour

Configuration Files:

/etc/hosts - Count: 1

Security:

Unsigned Files:

Launchd: /Library/LaunchAgents/com.oracle.java.Java-Updater.plist

Executable: /Library/Internet Plug-Ins/JavaAppletPlugin.plugin/Contents/Resources/Java Updater.app/Contents/MacOS/Java Updater -bgcheck

Details: Exact match found in the whitelist - probably OK

Launchd: /Library/LaunchDaemons/org.virtualbox.startup.plist

Executable: /Library/Application Support/VirtualBox/LaunchDaemons/VirtualBoxStartup.sh restart

Details: Exact match found in the whitelist - probably OK

Launchd: /Library/LaunchDaemons/com.oracle.java.Helper-Tool.plist

Executable: /Library/Internet Plug-Ins/JavaAppletPlugin.plugin/Contents/Resources/Helper-Tool

Details: Exact match found in the whitelist - probably OK

32-bit Applications:

11 32-bit apps

Kernel Extensions:

/Library/Application Support/VirtualBox

VBoxDrv.kext (Oracle America, Inc., 5.2.18)

VBoxNetAdp.kext (Oracle America, Inc., 5.2.18)

VBoxNetFlt.kext (Oracle America, Inc., 5.2.18)

VBoxUSB.kext (Oracle America, Inc., 5.2.18)

/Library/Extensions

ufsd_NTFS.kext (Paragon Software GmbH, 15.2.319 - SDK 10.10)

TACC.kext (TechSmith Corporation, 1.0.3 - SDK 10.10)

/System/Library/Extensions

MotMobileUSB.kext (1.2.2 - SDK 10.5)

/System/Library/Extensions/MotMobileUSB.kext/Contents/PlugIns

MotMobileMS.kext (1.0.0 - SDK 10.5)

MotMobileMTP.kext (1.2.2 - SDK 10.5)

MotMobileUSBLAN.kext (1.2.2 - SDK 10.5)

MotMobileUSBLANMerge.kext (1.2.2 - SDK 10.5)

MotMobileUSBSwch.kext (1.2.2 - SDK 10.5)

System Launch Agents:

System Launch Daemons:

Launch Agents:

Launch Daemons:

Internet Plug-ins:

JavaAppletPlugin: Java 8 Update 181 build 13 (installed 2018-08-03)

AdobeAAMDetect: 3.0.0.0 (installed 2018-08-01)

QuickTime Plugin: 7.7.3 (installed 2018-08-20)

3rd Party Preference Panes:

Java (installed 2018-08-03)

NTFS (installed 2018-04-20)

Performance:

System Load: 1.62 (1 min ago) 2.98 (5 min ago) 3.03 (15 min ago)

Nominal I/O speed: 4.34 MB/s

File system: 24.12 seconds

Write speed: 1482 MB/s

Read speed: 2014 MB/s

CPU Usage:

Top Processes by CPU:

Top Processes by Memory:

Top Processes by Network Use:

Virtual Memory Information:

Software Installs (past 30 days):

End of report

You have several kernel extensions that could cause problems:

/Library/Extensions

ufsd_NTFS.kext (Paragon Software GmbH, 15.2.319 - SDK 10.10)

TACC.kext (TechSmith Corporation, 1.0.3 - SDK 10.10)

/System/Library/Extensions

MotMobileUSB.kext (1.2.2 - SDK 10.5)

You have also run Etrecheck in a mode that hides what processes were running (possibly Sandbox mode). Was kernel_task running away when you ran Etrecheck? It doesn't appear so, but I can't tell. If it wasn't, then you should run it when you see kernel_task running away.

Beta version in the wild! The current version of EtreCheck doesn't work in Mojave, so I direct Mojave users to the beta. After all, what's another beta?

If this sandboxed version of EtreCheck detected any out of control kernel_task usage, it would show up under "Other processes". But it would have to be a huge portion of the CPU. That machine is mostly idle. The "kernel_task" problem is pretty rare and even more rare to catch it with EtreCheck, especially the beta version. The one example I have seen showed all CPU cores with very heavy system CPU usage. This included the virtual cores, which are even other one in the list, after the overall CPU usage. I've never seen virtual cores have any significant usage otherwise.

With High Sierra, I have seen a few instances where there can be significant kernel_task usage. It can go up to 100% of slightly higher. That is normal. In my case, it appeared to be due to an APFS snapshot being generated. This would show up as a fairly high level of "Other processes" with a fairly high level of "Nominal I/O speed". This particular report could be in that range. If so, it is nothing to worry about. The actual "kernel_task" problem would have much higher CPU usage and would also throttle disk I/O.

I had a similar problem with my 15" MacBook Pro Early 2011 running macOS High Sierra (10.13.6). I tried restoring the OS, among other things, to no avail. I noticed that the problem occurred much earlier in the recent hot weather here. I put the laptop on a ice pack and was able to demonstrate that it worked very well when cooled, and bogged down in kernal_task when allowed to warm up - completely controllable.

So, i removed the base. There was dust, but it did not seem like much. I used a strong houshold vacuum cleaner to carefully vacuum the visible dust and I also vacuumed the rear vents (causing the fans to spin rapidly) - it is possible that there was a lot of dust between the motherboard and keyboard. After re-assembling, i also vacuumed the keyboard and speaker ports. It appears to me that the fans exhaust hot air to the rear but the cold air inlet openings are not obvious - possibly the keyboard or speaker vents?

Anyway, it magically resolved the issues - the MacBook Pro behaves like it did when new and never bogs down with kernal_task, even under a high CPU load.

Because I rarely use the MacBook Pro as a laptop - it is more or less permanently hooked up to a Thunderbolt display - I have been playing with a cooling pad (Aluminum grate with fans) and it works, keeping the temperatures even lower (I have a iStat widget that reports temperature of 3 enclosure locations, CPU, CPU heat sink and internal disk) - the Aluminum grate has heat sink benefit even without the external fans running. I have no evidence that this supplemental cooling is required, however. Probably periodic vacuuming will keep it running well enough.