I do agree on that, and as was stated before on the list, the load is a
systemwide number, on the aerage lenghht on the running queue, and not
processor specific.
Some people however think of `load' as the actual usage of the cpu-cycles
as the `load' on one processor. That's why I mentioned the procps-util
pack I found somewhere.
-- Merijn
On Sun, 28 Nov 1999, Mark Garlanger wrote:
> Date: Sun, 28 Nov 1999 17:27:57 +0000
> From: Mark Garlanger <[EMAIL PROTECTED]>
> To: Merijn Vogel <[EMAIL PROTECTED]>
> Cc: [EMAIL PROTECTED]
> Subject: Re: SMP Linux processor loads
>
> That's nice, but that wasn't the question. 'cpu-allocation for the processes' does
>NOT
> mean LOAD.
>
> Mark
>
> Merijn Vogel wrote:
>
> > There is a top and ps (which shows cpu-allocation for the processes).
> > Can't find the url, one place to find it is at my homepage:
> >
> > http://web.inter.nl.net/users/merijnv/procps-1.2.9-SMP-GW.patched.tgz
> >
> > So long,
> >
> > -- Merijn
> >
> > On Sat, 27 Nov 1999, Mark Garlanger wrote:
> >
> > > Date: Sat, 27 Nov 1999 19:09:10 +0000
> > > From: Mark Garlanger <[EMAIL PROTECTED]>
> > > To: Darren Nickerson <[EMAIL PROTECTED]>
> > > Cc: [EMAIL PROTECTED], Juan Jose Casero <[EMAIL PROTECTED]>,
> > > [EMAIL PROTECTED], [EMAIL PROTECTED]
> > > Subject: Re: SMP Linux processor loads
> > >
> > > Darren Nickerson wrote:
> > >
> > > > >
> > > > >
> > > > > On Fri, 26 Nov 1999, Juan Jose Casero wrote:
> > > > >
> > > > > > Hi Folks -
> > > > > >
> > > > > > Is there a way to observe the load on each cpu in an SMP linux system with
> > > > > > kernel 2.2.13? I am curious to see how the linux kernel distributes
>system
> > > > > > resources to processes. Additionally, how does Linux smp compare with say
> > > > > > NT smp or even with Solaris x86 smp?
> > > > > >
> > > > > > Thanks.....
> > > > > >
> > > > > try top,gtop,or xosview
> > > > >
> > > > >
> > > > > I think linux is still lagging behind but is improving at a high rate.
> > > >
> > > > Justin, as far as I can tell none of these do what the original poster was
> > > > asking . . . ie, show the load on EACH cpu independently (at least in my distro
> > > > - redhat-6.1). Why are so few of the userspace tools SMP-aware yet?
> > >
> > > No tools do that, because that is not a valid statistic. The load is NOT how
>busy
> > > a given processor is. It is
> > > how many processes are ready to run. Unless you do some non-standard kernal hack
> > > to allow you
> > > to partition processes to a given processor (or a possible group of processors
>if
> > > you have more the
> > > 2 CPUs), there is only a single queue for the system. Thus only one load for the
> > > system.
> > >
> > >
> > > Mark
> > >
> > > -
> > > Linux SMP list: FIRST see FAQ at http://www.irisa.fr/prive/dmentre/smp-howto/
> > > To Unsubscribe: send "unsubscribe linux-smp" to [EMAIL PROTECTED]
> > >
> >
> > -- - Strong encryption for everyone. Get my public GPG key at:
> > -- http://web.inter.NL.net/users/merijnv/publickey.asc or at a keyserver.
> > -- http://www.keyserver.net/
> > -- Homepage: http://web.inter.NL.net/users/merijnv/
> >
> > -
> > Linux SMP list: FIRST see FAQ at http://www.irisa.fr/prive/dmentre/smp-howto/
> > To Unsubscribe: send "unsubscribe linux-smp" to [EMAIL PROTECTED]
>
> -
> Linux SMP list: FIRST see FAQ at http://www.irisa.fr/prive/dmentre/smp-howto/
> To Unsubscribe: send "unsubscribe linux-smp" to [EMAIL PROTECTED]
>
-- - Strong encryption for everyone. Get my public GPG key at:
-- http://web.inter.NL.net/users/merijnv/publickey.asc or at a keyserver.
-- http://www.keyserver.net/
-- Homepage: http://web.inter.NL.net/users/merijnv/
-
Linux SMP list: FIRST see FAQ at http://www.irisa.fr/prive/dmentre/smp-howto/
To Unsubscribe: send "unsubscribe linux-smp" to [EMAIL PROTECTED]
