Subject: Re: One data point: PC/Su From: timo.suhonen@lill.frmug.fr.mugnet.org (Timo Suhonen) Date: Tue, 1 Jun 1993 10:02:00 +0001
ÿ@SUBJECT:Re: One data point: PC/Sun == 1/6
ÿ@ORIGIN :University of Jyvaskyla, Department of Biology of Physical
N
Message-ID: <SUHONEN.93Jun1120248@kunto.jyu.fi>
Newsgroup: comp.os.linux,comp.unix.pc-clone.32bit,comp.sys.sun.misc
malouf@leland.Stanford.EDU (Rob Malouf) writes:
The first time I compiled the Linux kernel on my 4 meg 486-33DX, it
took ~2.5 hours. After I went back to the vendor to get the
motherboard replaced so the cache would work, in took 1.25 hours.
Last week, I upgraded to 8 meg, and it took 16 minutes. Clearly,
excessive page faulting greatly reduces the speed of gcc.
Hmm... I have started to think that I'm doing something totally wrong:
I have a 386/7-25 with 5 meg and can compile the kernel under 30 minutes.
And that is with the old filesystem (minix-based). My disk is a 3 year
old Nec... Maybe this one extra meg (from 4 to 5) is the key???