Date: Thu, 2 May 2002 18:30:56 -0700 (PDT) From: Patrick Thomas <root@utility.clubscholarship.com> To: Walter Hop <walter@binity.com> Cc: <freebsd-questions@freebsd.org> Subject: Re: confusing CPU stats in `top` Message-ID: <20020502182856.F29785-100000@utility.clubscholarship.com> In-Reply-To: <20020503011942.N67241-100000@surreal.nl>
next in thread | previous in thread | raw e-mail | index | archive | help
Does this mean I should be less worried about a quad-proc machine at overall 50% idle than a single processor machine at 50% idle ? (because a single proc system at 50% idle worries me....) thanks, PT On Fri, 3 May 2002, Walter Hop wrote: > [in reply to Patrick Thomas <root@utility.clubscholarship.com>, 02/05/02] > > > As you can see, the server reports 69.7% idle > > > > BUT, that would imply processes using up a total of (30.3 x 4) `top` > > reported percent ... which is not the case - if you add up the > > percentages in the `top` reported processes, they add up to ~58 or so, > > and if we divide that by four we would expect to see 85% idle or so, > > instead of 69.7% idle. > > The FreeBSD 4.x kernel has an inefficiency in SMP systems: while one > processor is running in the kernel, the other processors cannot due to > resource locking, so they have to wait. This causes the system to be > less idle than an ideal system. > > (FreeBSD 5 is expected to solve a lot of these issues, but also with > that kernel you will see that you can't just multiply CPU power by the > number of processors, there's always some overhead involved.) > > Also the numbers in "top" are just rough measures.. > > -- > Walter Hop <walter@binity.com> | +31 6 24452020 | PGP keyid 0x84813998 > > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020502182856.F29785-100000>