Date: Fri, 26 Nov 2004 11:44:53 +0900 From: Rob <spamrefuse@yahoo.com> To: freebsd-stable@freebsd.org Subject: Re: make -j$n buildworld : use of -j investigated Message-ID: <41A698A5.6010601@yahoo.com> In-Reply-To: <3282.1101377849@thrush.ravenbrook.com> References: <opsh0jp9ge8527sy@smtp.local> <3282.1101377849@thrush.ravenbrook.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Nick Barnes wrote:
> On Thu, 25 Nov 2004 16:19:02 +0900, Rob <spamrefuse@yahoo.com> wrote:
>
>
>>> time(minutes) * speed(MHz) * nproc / 1000 MHz
>
>
> Looking at your examples, it seems you divide by 1e5, not by 1000.
Sorry, yes you're right.
> In other words, buildworld is CPU bound and takes about 6e12 clock
> cycles. Use -j<nproc>.
My main point is: when using "-j<nproc>", I am able to predict how long
the buildworld takes on a range of 'mainstream' systems like the ones I
have (pentium-II to pentium-IV and dual-pentium-III):
time(minutes) = 1e5 / ( speed(MHz) * nproc )
P-II 333 MHz : 300 min.
P-III 700 MHz : 142 min.
P-III 2x800 MHz : 62 min.
P-IV 2000 MHz : 50 min.
Within a few minutes, all these times are correct.
I still have to test my P-I/150 MHz also running 5.3, and verify
that buildworld will take approx. 670 min.
Rob.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?41A698A5.6010601>