Star InactiveStar InactiveStar InactiveStar InactiveStar Inactive
 

We now have a new Raspberry4 gibt, which in contrast to the previous versions becomes quite hot. So everybody is interested in the effectiveness of either passive or active cooling and whether this cooling suppresses CPU throtteling.

I wrote a small script check_throttled.sh which reads the current throttling states of the Raspberry with vcgencmd get_throttled and if throtteling happens or already happend the meaning of the throttling bits is reported.

 

 

 

Example output check_throttled.sh:

pi@raspberrypi-buster:~ $ ./check_throttled.sh

Throttling in hex (bits reset on boot): 0x20000

Bit 17 set: Arm frequency capped has occurred

Throttling in hex: 0x20002 (bits reset every call)

You can see the Raspberry was throttled at least once since reboot but currently there is no throtteling.

 

In addition i Wrote a small script temp_test.sh which allows to determine the max CPU temperature which a Raspberry reaches if there is a 100% CPU utilization. In addition you can display the increasing CPU temperature in a configurable interval.

 

Example output temp_test.sh:

pi@raspberrypi-buster:~ $ ./temp_test.sh -i 5

Generate 100% CPU utilization and measure CPU temperature ...

CPU watch interval: 5s

  1. Watch +0s:temp=55.8'C
  2. Starting run 1: +0s:temp=56.4'C
  3. Watch +5s:temp=64.5'C
  4. Watch +10s:temp=67.7'C
  5. Watch +15s:temp=70.4'C
  6. Watch +20s:temp=73.1'C
  7. Watch +25s:temp=74.1'C
  8. Watch +30s:temp=76.3'C
  9. Watch +35s:temp=77.4'C
  10. Watch +40s:temp=79.0'C
  11. Watch +45s:temp=80.6'C
  12. Watch +50s:temp=80.6'C
  13. Watch +55s:temp=81.1'C
  14. Watch +60s:temp=81.7'C
  15. Watch +65s:temp=82.2'C
  16. Watch +70s:temp=82.7'C
  17. Watch +75s:temp=82.7'C

 

References:

1) Reset of throtteling values

2) Meaning of throtteling values

Add comment

Note

Spam comments are purged and not published. Review is done manually and therefore it usually takes between some hours and one day until a comment will be published.