User:Jukeboksi/Blog/2018: Difference between revisions

→‎Tuesday 2018-10-23: UPDATE: tested another computer and its power consumption grew from 17W (fully idle) to 52W (fully stressed CPU) so that is a sizable +206% increase
(Today I did a small test to find out how much CPU load affects the power consumption of a laptop.... the result of this test is that my laptop uses +35% more energy when the CPU cores are under a full workload, compared to the computer mostly idling.)
(→‎Tuesday 2018-10-23: UPDATE: tested another computer and its power consumption grew from 17W (fully idle) to 52W (fully stressed CPU) so that is a sizable +206% increase)
Line 2: Line 2:


=== Tuesday 2018-10-23 ===
=== Tuesday 2018-10-23 ===
* Today I did a small test to find out how much CPU load affects the power consumption of a laptop. I did this because I have recently resumed contributing my computational resources to the [[w:Berkeley Open Infrastructure for Network Computing]] or BOINC for short and am considering whether to add running BOINC to the things suggested to consumers who stumble upon the development wiki [[Main Page]]. When the computer, a Lenovo ThinkPad X230, was being '''mostly idle''' it consumed on average '''30 watts''' or slightly under and when I put all cores into full labour with the [[w:hardware stress test]] program for Linux aptly named just 'stress' with the parameter '--cpu 4' (the CPU has 4 cores), which causes CPU utilization going to 100%. '''Under full workload''' on the CPU cores the computer used an average of '''40 watts'''. Based on this sample of one laptop the result would seem to be that '''cranking all CPU cores to the max''' causes an approximate '''+35% electricity consumption'''. [[User:Jukeboksi|Jukeboksi]] ([[User talk:Jukeboksi|talk]]) 13:23, 23 October 2018 (UTC)
* Today I did a small test to find out how much CPU load affects the power consumption of a laptop. I did this because I have recently resumed contributing my computational resources to the [[w:Berkeley Open Infrastructure for Network Computing]] or BOINC for short and am considering whether to add running BOINC to the things suggested to consumers who stumble upon the development wiki [[Main Page]].  
 
:When, a Lenovo ThinkPad X230, was being '''mostly idle''' at 20-30% CPU utilization it consumed on average '''30W''' or slightly under and when I put all cores into full labour with the [[w:hardware stress test]] program for Linux aptly named just 'stress' with the parameter '--cpu 4' (the CPU has 4 cores), which causes CPU utilization going to 100%. '''Under full workload''' on the CPU cores the computer used an average of '''40W''' so  '''cranking all CPU cores to the max''' causes an approximate '''+35% electricity consumption'''.
 
:Also tested an another computer, a Dell Inpiron E6220 that had a very low power consumption of around '''17W''' when idle with 10% of CPU in use and the stressing of the CPU to the max caused a much bigger rise in its electricity consumption: It went to '''52W''' so that is a rise of '''+206% electricity consumption''' or '''+35W''' in terms of absolute consumption.
 
:It is hard to draw a conclusion from a sample of only 2 computers, but it is clear that CPU utilization does play a major role in how much power the computer consumes. [[User:Jukeboksi|Jukeboksi]] ([[User talk:Jukeboksi|talk]]) 13:41, 23 October 2018 (UTC)


=== Monday 2018-10-15 ===
=== Monday 2018-10-15 ===
9,854

edits