Find your own processor's limits and watch it throttle
Task
Establish what your processor actually is, what its thermal design power is, and then make it reach its limit under load so that throttling is something you have observed rather than read about.
Steps
- Record the processor's exact model, socket, core and thread count, base and boost clocks, and thermal design power in
lab/cpu/spec.csvwith headerfield,value,source--sourcenames the manufacturer page, not a guess. - Record idle temperature and clock speed, averaged over one minute, in
lab/cpu/idle.csvwith headerreading,temp_c,clock_mhz. - Run a processor stress test for ten minutes, logging temperature and clock speed at least every thirty seconds into
lab/cpu/load.csvwith the same header. - Write
lab/cpu/observed.mdstating the peak temperature, whether the clock fell below its sustained boost, and at what temperature that began. - Write
lab/cpu/cooling.mdlisting the three things you would check first if the peak temperature had been within five degrees of the processor's stated maximum.
Verify
awk -F, 'NR>1 && NF>=3 {n++} END {print n" specification field(s)"}' lab/cpu/spec.csv
awk -F, 'NR>1 && NF>=3 {n++} END {print n" idle reading(s)"}' lab/cpu/idle.csv
awk -F, 'NR>1 && NF>=3 {n++} END {print n" load reading(s)"}' lab/cpu/load.csv
awk -F, 'NR>1 && $2>0 {if($2>m)m=$2} END {print "peak "m" C"}' lab/cpu/load.csv
grep -c . lab/cpu/observed.md
grep -c . lab/cpu/cooling.md
Five or more specification fields with sources, twenty or more load readings, a peak temperature computed from your own data, and both notes present. If the clock never varied across twenty readings, either the cooling has ample headroom or the stress test did not load the processor.
Notes
Stop the test if the temperature reaches the processor's stated maximum. That is the machine protecting itself and there is nothing further to learn from holding it there.
This is an independent study companion for CompTIA A+ Core 1 220-1201 and is not produced by or endorsed by CompTIA.