[ltp] Re: Stress testing for undervolting

[Laurent Gilson]

Hello,

> I think I can figure out to do the first and third of your suggestions,  
> but
> I'm not exactly versed in all of hdparm's options, and even after reading
> the manpage, I'm not sure what switches you have in mind for a  
> random-seek
> test.

Err..., sorry my fault, wrong prog. i donīt know what i have used to make  
that test. i guess one of these:  
http://people.redhat.com/dledford/benchmark.html

>> I have tested 2 hours in total. And added 1 voltage step afterwards.  
>> After
>> 4 months: No hiccups.
>
> Was this 2 hours of straight mprime, or mprime augmented with your more
> fanciful ideas above?

First i searched the absolute minimum voltage for each frequency by  
running mprime and decreasing the voltage by 1 step every 10 seconds. That  
way i found the drop-off points (and crashed/rebooted 2-3 times).  
Afterwards i increased every voltage by 1 step (2 for the ones that  
crashed hard).

Then 2 hours torture test with all the things i wrote in the first mail  
(mprime, seeks, WLAN, changing the frequency via script every 30 sec). No  
crashs, no error in mprime. But it still didnīt "feel" safe so i added  
another step to each voltage.

After 3 hours surfing/office my CPU is at 42-45°C, w/o any fan running. I  
have 3:45 runtime from a 40Wh battery (with WLAN switched on but bluetooth  
off) (IBM promises 3:00 for a 47Wh... go figure). I didnīt bother to check  
if the last "safety" step was really necessary.

cu