top / cpu usage and cpu freq scaling?
in [Suse]
|
Prev: eSATA issues
Next: hyper terminal
From: jamm on 17 Mar 2010 18:25 hmm I wonder, how does a cpu usage measuing utility like top factor in the current CPU frequency? Or does it?
From: DenverD on 18 Mar 2010 02:24 jamm wrote: > hmm I wonder, how does a cpu usage measuing utility like top factor in the > current CPU frequency? Or does it? if it does, it seems that neither google nor the easy to find documentation (man and info) for top speaks of it.. as far as i can find, that is.. but, i'm sure you looked there before you asked here, so i have to assume you *must* have found something that caused you to continue to wonder.. what was it? maybe you emailed top's author, and he said....what? -- DenverD (Linux Counter 282315) via Thunderbird 2.0.0.23 (20090817), KDE 3.5.7 "release 72-11", openSUSE Linux 10.3, 2.6.22.19-0.4-default #1 SMP i686 athlon
From: jamm on 18 Mar 2010 03:17 DenverD wrote: > jamm wrote: >> hmm I wonder, how does a cpu usage measuing utility like top factor in >> the current CPU frequency? Or does it? > > if it does, it seems that neither google nor the easy to find > documentation (man and info) for top speaks of it.. > > as far as i can find, that is.. > > but, i'm sure you looked there before you asked here, so i have to > assume you *must* have found something that caused you to continue to > wonder.. > > what was it? > It just occured to me as I dropped a cpu freq plasmoid onto my desktop.. yes I did try to scroggle it but found nothing. I have a 3 Ghz processor. When its throttled to 1 Ghz, 100% cpu usage would really be 33% cpu usage at 3Ghz... so it should show 33% right? I doubt it works that way though. They probably kept it simple with no taking account of changing freq. just curiosity :)
From: Will Honea on 18 Mar 2010 18:11 jamm wrote: > DenverD wrote: > >> jamm wrote: >>> hmm I wonder, how does a cpu usage measuing utility like top factor in >>> the current CPU frequency? Or does it? >> >> if it does, it seems that neither google nor the easy to find >> documentation (man and info) for top speaks of it.. >> >> as far as i can find, that is.. >> >> but, i'm sure you looked there before you asked here, so i have to >> assume you *must* have found something that caused you to continue to >> wonder.. >> >> what was it? >> > > It just occured to me as I dropped a cpu freq plasmoid onto my desktop.. > yes I did try to scroggle it but found nothing. > > I have a 3 Ghz processor. When its throttled to 1 Ghz, 100% cpu usage > would really be 33% cpu usage at 3Ghz... so it should show 33% right? I > doubt it works that way though. They probably kept it simple with no > taking account of changing freq. > > just curiosity :) While that sounds good, most of the CPU usage monitors I've worked with simply measure the amount of idle time vs the active time and make no reference to the actual frequency vs. potential frequency. Thus 100% at 1 Ghz is the same as 100% at 3 Ghz - 100% of the CPU cycles available at the current clock rate. Of course, that still means that you only get 1/3 the CPU cycles in the measured time span. It also means that the speed scheduling is not doing its' job - 100% at a lower clock rate should almost never happen as a proper algorithm should increase the speed (thus cycles available) well before 100% is reached. I think my single core laptop bumps the speed at something like 50% usage. -- Will Honea
From: Paul J Gans on 19 Mar 2010 21:41
jamm <nospam(a)nomail.net> wrote: >DenverD wrote: >> jamm wrote: >>> hmm I wonder, how does a cpu usage measuing utility like top factor in >>> the current CPU frequency? Or does it? >> >> if it does, it seems that neither google nor the easy to find >> documentation (man and info) for top speaks of it.. >> >> as far as i can find, that is.. >> >> but, i'm sure you looked there before you asked here, so i have to >> assume you *must* have found something that caused you to continue to >> wonder.. >> >> what was it? >> >It just occured to me as I dropped a cpu freq plasmoid onto my desktop.. yes >I did try to scroggle it but found nothing. >I have a 3 Ghz processor. When its throttled to 1 Ghz, 100% cpu usage would >really be 33% cpu usage at 3Ghz... so it should show 33% right? I doubt it >works that way though. They probably kept it simple with no taking account >of changing freq. >just curiosity :) I thought that cpu usage reported, approximately, the percent of current capacity in use. Speed has to do with how long a given process will run. Cpu capacity, in my opinion, doesn't measure that, nor should it. -- --- Paul J. Gans |