well done !
Phoronix: Nailing Down The Linux Kernel Power Regressions
For those that missed it, now that I've been back in the office for a few weeks and caught up on the other benchmarks and articles for May and beginning the work for June, I'm back to looking into the Linux kernel power regressions reported more than a month ago -- i.e. Linux Has Major Power Regression and Another Major Linux Power Regression Spotted...
well done !
I thought this was traced back a while ago?
Have you tried creating a cut down bare bones kernel? This should make life much easier and quicker for bisecting over a large number of versions. Of course you'd need to check the problem still exists first in the cut down kernel
If you need help with this Michael please give me a shout
Good to see it's not forgotten. Please keep up the good work!
Maybe IBM, Google, Oracle and co could be interested in supporting you? ; )
are you using ccache or distcc.
How about modifying a worn out battery by completely removing its cells and connecting a power supply where the cells had been connected? Then do something like:... still needing to monitor them for manually toggling the AC power / battery. With that said, I'm still looking for a UPS power system (or a smart AC power meter) with ...
grep "present rate:" /proc/acpi/battery/BAT0/state
grep "present voltage:" /proc/acpi/battery/BAT0/state
Warning: Won't eat your cat but might otherwise be harmful (or might not work at all).
Edit: Please check that number of cells and the voltage of power supply matches
Last edited by frief; 06-01-2011 at 06:45 PM. Reason: added caveat
Just my idea:
Point a web-cam to this device, do a little OCR magic and you have your numbers.
I just stumbled across a patch for (soon to arrive) version 22.214.171.124:
Maybe this small patch solves the power regression.
I'm currently building a vanilla 2.6.39 kernel with this patch applied. Tomorrow i will compare it with a completely vanilla kernel. Maybe the power consumption is significantly lower with this patch.
In order to measure power consumption I'd suggest to use multimeters instead of an UPS. The precision of the UPS is not known and it can change greatly from a manufacturer to another leading to unreproducible results.
A bit over your budget ($140 to $150) but a better solution would be to use two VA18B DMM with usb interface, one to measure voltage, the other to measure current and then combine the two to have power.
At http://www.multimeterwarehouse.com/VA18B.htm the VA18 is $70 and at http://multimeter.schewe.com/ you can get a software to read data from usb.
It involves a little "surgery" on the power chord in order to measure the current (you must cut one cable and attach 2 banana plugs to put the amper-meter serially) but the measures will be better then with a reading from a UPS.
If needed I can provide a more detailed schematic.