From: agc@weasel.demon.co.uk (Alan Charlton) Subject: Re: Kernel build benchmark Date: 22 Dec 1992 10:48:08 GMT
In article <BzM0HC.n39@citrus.SAC.CA.US> ianj@citrus.SAC.CA.US ( Ian Justman ) writes:
> cae@davidge.com (Caleb Epstein) writes:
> : A while back I recall a number of people posting Linux
> : 'benchmarks' based on how long it took their machine to
> : rebuild the kernel. The less time it took, the faster the
> : machine.
> Uh-uh... That sounds like a good analogy; however, what about if
> you had more memory like I do (I have 3MB)?
>
> I'm using a 386sx/20 with, like I said, 3MB of RAM and 5 megs'
> worth of swap partition, and gee whiz, I thought _ONE_ HOUR was
> bad! Evedently even one megabyte of RAM makes all the diff (no
> pun intended... Ok, Ok, OK!!! Pun intended :-) ).
>
> : This little UNIX is a tribute to its author(s). The
> : fact that I can compile *anything* on this hunk-of-junk PC
> : is near miraculous.
>
> I second the emotion!
>
> Here's to Linus B. Torvalds and his masterpiece which only gets
> better!
>
I'm using a 386sx/25 with 8Mb of RAM and 3M swap (of which only about 512K
is ever used), and a 0.99 make dep/make Image takes about 40-50 minutes,
and never more than 1 hour. It would seem that lots of memory has a far
greater influence on build times than all-out processor speed.
Cheers,
Alan.
-- +---------------------------------+--------------------------------+ | Alan Charlton, Bristol, England | If B.U.R.D. doesn't spell Bird | | agc@cix.compulink.co.uk | - what the hell does it spell? | | agc@weasel.demon.co.uk | | | Tel: +44 454 202706 (after 6) | --- W. R. Hearst | +---------------------------------+--------------------------------+