Quantcast
Channel: Raspberry Pi Forums
Viewing all articles
Browse latest Browse all 8082

Advanced users • Re: Moving Linux Kernel to 6.18

$
0
0
Can i confirm you are not overclocking?
No overclocking (I always get the CPUs which will not overclock well. Same with this one)
What power supply are you using?
Official 27W plug
Do you have any cooling?
The official fan cooler for Pi5
Are you 100% sure that 6.12 kernel never has this issue (compared to it just being intermittent and "being lucky so far").
Yes

Latest test runs:
moved the pi next to the window for cooler air.
used a ugreen power bank with 140W. Shall deliver 60W on the used USB C port.
over_voltage=6
btop and small script (at end of this post) for monitoring.

result: Kernel 6.18 has problem with power/temperature handling.

6.18: starts with 1.5GHz and 1.00V. Keeps 1.00V when going to full 2.4GHz. Keeps under load all cores at 100%. When reaching temperature of 70C there is a high chance that the error occurs. Temps can get much higher (76C) when errors show up.

snippet from log file $TIMESTAMP; ${TEMP}; ${FREQ}; ${VOLT}; ${THROT}" :

Code:

2026-01-06 20:52:02; 32.9; 1500016128; 1.0000V; 0x02026-01-06 20:52:12; 31.8; 1500012800; 1.0000V; 0x02026-01-06 20:52:22; 32.9; 1500019456; 1.0000V; 0x02026-01-06 20:52:32; 32.9; 1500016128; 1.0000V; 0x02026-01-06 20:52:42; 33.4; 1500016128; 1.0000V; 0x02026-01-06 20:52:52; 32.3; 2400023808; 1.0000V; 0x02026-01-06 20:53:02; 34.5; 1700014464; 1.0000V; 0x02026-01-06 20:53:12; 35.1; 2400027136; 1.0000V; 0x02026-01-06 20:53:22; 51.6; 2400020480; 1.0000V; 0x02026-01-06 20:53:32; 53.2; 2400027136; 1.0000V; 0x02026-01-06 20:53:42; 56.0; 2400030464; 1.0000V; 0x02026-01-06 20:53:52; 45.0; 1500012800; 1.0000V; 0x02026-01-06 20:54:02; 44.4; 2400020480; 1.0000V; 0x0
6.12: start with 1.5GHz and 0.75V. Goes up to 1.00V when going to full 2.4GHZ. When reaching core temp of 70C 1-2 cores do not reach 100%. When temperature gets lower they reach 100% again.

Code:

2026-01-07 01:02:40; 42.8; 1500012800; 0.7500V; 0x02026-01-07 01:02:50; 41.1; 1500016128; 0.7500V; 0x02026-01-07 01:03:22; 41.1; 1500012800; 0.7500V; 0x02026-01-07 01:03:32; 40.0; 1500022656; 0.7500V; 0x02026-01-07 01:03:42; 39.5; 1500016128; 0.7500V; 0x02026-01-07 01:03:52; 40.0; 1500012800; 0.7500V; 0x02026-01-07 01:04:02; 41.1; 1700017792; 0.9480V; 0x02026-01-07 01:04:12; 41.1; 2400033792; 1.0000V; 0x02026-01-07 01:04:23; 54.9; 2400033792; 1.0000V; 0x02026-01-07 01:04:33; 59.3; 2400017408; 1.0000V; 0x02026-01-07 01:10:44; 73.6; 2400027136; 1.0000V; 0x02026-01-07 01:10:54; 73.6; 2400027136; 1.0000V; 0x02026-01-07 01:11:04; 57.6; 1500019456; 0.7500V; 0x02026-01-07 01:11:14; 53.8; 1500016128; 0.7500V; 0x02026-01-07 01:11:24; 51.0; 1500019456; 0.7500V; 0x02026-01-07 01:11:34; 49.9; 1500022656; 0.7500V; 0x02026-01-07 01:11:44; 49.4; 1500012800; 0.7500V; 0x02026-01-07 01:11:54; 64.8; 2400020480; 1.0000V; 0x02026-01-07 01:12:04; 67.0; 2400020480; 1.0000V; 0x02026-01-07 01:12:14; 67.5; 2400030464; 1.0000V; 0x0
With both kernels I did not reach 80C, so get_throttled is allways 0x0

Monitoring script:

Code:

#!/bin/bashLOGFILE="./temperature_log.txt"INTERVAL=10  # Log every 10 secondsecho "Logging CPU temperature to $LOGFILE"while true; doTIMESTAMP=$(date +"%Y-%m-%d %H:%M:%S")TTT=$(vcgencmd measure_temp | egrep -o '[0-9]*\.[0-9]*')FREQ=$(vcgencmd measure_clock arm | egrep  -o '([0-9]+$)')VOLT=$(vcgencmd measure_volts | egrep  -o '([0-9.]+V)')THROT=$(vcgencmd get_throttled | egrep -o '([0-9a-fA-Fx]+)$')echo "$TIMESTAMP; ${TTT}; ${FREQ}; ${VOLT}; ${THROT}" >> "$LOGFILE"echo "$TIMESTAMP; ${TTT}; ${FREQ}; ${VOLT}; ${THROT}" sleep "$INTERVAL"done

Statistics: Posted by crumble — Wed Jan 07, 2026 1:40 am



Viewing all articles
Browse latest Browse all 8082

Trending Articles