Re: [Powertop] How does powertop deal with multiple executions
Jesús D <jesus.diaz.vico at gmail.com>
| Newsgroups | dev.linux.lists.powertop |
|---|---|
| Message-ID | <CAA3v608roAC6tWo6DQ+NU6fyvptFw_weCw=PWhsykeqrfVYo1w@mail.gmail.com> |
Thank you, Arjan. But my question is how does powertop group programs with the same name. Unless I'm doing something wrong, I doubt that it is just adding the power consumption of all the independent executions. I say this because powertop estimates a consumption of 335 mW for 500 executions, and a consumption of 364 mW for 5000. I am configuring the monitorization time with the --time parameter (with a value long enough to cover all the 500 or 5000 executions). Is it calculating an average then? On Tue, Oct 29, 2013 at 3:20 AM, Arjan van de Ven <arjan(a)linux.intel.com>wrote: > On 10/28/2013 4:37 PM, Jesús Díaz Vico wrote: > >> Hello, >> >> I'm trying to estimate the power consumption (besides some other >> properties) of a custom program. During my tests, I execute my program >> several times (it is not a >> background process) with powertop running. My doubt is: how does powertop >> deal with these multiple executions? >> >> I'd guess that it averages over the number of executions, because I'm >> obtaining very similar power consumption measurements for both 500 and 5000 >> executions of my program >> (335 mW and 364 mW, respectively). Still, there is the possibility that >> powertop only keeps the estimation corresponding to the last execution... >> >> Thanks. >> > > by default, powertop will group programs with the same name into one > bucket for display purposes. > > > -- Jesús Díaz Vico
attachment.html
(text/html, 1.9 KB)
<div dir="ltr"><div>Thank you, Arjan.<br><br>But my question is how does powertop group programs with the same name. <br><br>Unless I'm doing something wrong, I doubt that it is just adding the power consumption of all the independent executions. I say this because powertop estimates a consumption of 335 mW for 500 executions, and a consumption of 364 mW for 5000. I am configuring the monitorization time with the --time parameter (with a value long enough to cover all the 500 or 5000 executions).<br> <br></div><div>Is it calculating an average then?<br></div></div><div class="gmail_extra"><br><br><div class="gmail_quote">On Tue, Oct 29, 2013 at 3:20 AM, Arjan van de Ven <span dir="ltr"><<a href="mailto:[email protected]" target="_blank">[email protected]</a>></span> wrote:<br> <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="HOEnZb"><div class="h5">On 10/28/2013 4:37 PM, Jesús Díaz Vico wrote:<br> <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> Hello,<br> <br> I'm trying to estimate the power consumption (besides some other properties) of a custom program. During my tests, I execute my program several times (it is not a<br> background process) with powertop running. My doubt is: how does powertop deal with these multiple executions?<br> <br> I'd guess that it averages over the number of executions, because I'm obtaining very similar power consumption measurements for both 500 and 5000 executions of my program<br> (335 mW and 364 mW, respectively). Still, there is the possibility that powertop only keeps the estimation corresponding to the last execution...<br> <br> Thanks.<br> </blockquote> <br></div></div> by default, powertop will group programs with the same name into one bucket for display purposes.<br> <br> <br> </blockquote></div><br><br clear="all"><br>-- <br>Jesús Díaz Vico </div>