Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 22 Jan 2002 15:44:53 -0800
From:      Kent Stewart <kstewart@owt.com>
To:        Greg Rumple <grumple@zaphon.llamas.net>
Cc:        freebsd-stable@FreeBSD.ORG
Subject:   Re: make -j4 world
Message-ID:  <3C4DF975.8050308@owt.com>
References:  <5.1.0.14.0.20020122145124.03cd1380@marble.sentex.ca> <5.1.0.14.0.20020122145124.03cd1380@marble.sentex.ca> <5.1.0.14.0.20020122151930.03d62220@marble.sentex.ca> <3C4DCB91.7080701@owt.com> <3C4DCDE7.C5E77C03@math.missouri.edu> <20020122125645.1a0c9b36.sreese@codysbooks.com> <20020122140431.R69194@zaphon.llamas.net>

next in thread | previous in thread | raw e-mail | index | archive | help


Greg Rumple wrote:

> * Scott Reese (sreese@codysbooks.com) [020122 20:52]:
> 
>>On Tue, 22 Jan 2002 14:39:03 -0600
>>Stephen Montgomery-Smith <stephen@math.missouri.edu> wrote:
>>
>>
>>>On the "Error in make buildworld" thread I said that I had this problem
>>>on 1 out of 3 machines, so you may not necessarily see this problem.  I
>>>only saw this problem on the fastest machine.
>>>
>>>
>>I've actually found that on a fast computer (1.6 GHz Pentium IV) using the "-j"
>>flag tends to slow things down a bit.  The standard "make buildworld" ran much
>>faster and much more smoothly than "make -j4 buildworld."  It was the same with
>>kernel builds.  Honestly, with a machine that fast, I don't really see a need
>>for the "-j" flags anyway.
>>
> 
> Well the machines I use make -j on are SMP machines, and using -j does
> make a difference.  I tend to use make -j 16 (2 p3-1Ghz cpu's, 2 gigs of
> ram) in order to finish as rapidly as possible.  This is significantly
> faster than make by it's self.  I have actually done a spreadsheet using
> this machine about 6 months ago, I should do it again.  Looking at the
> benefit of make -j finding the optimial value.  I haven't done a CVSUP
> up today, so not sure if this is happening to me or not.


My 4 runs with different -j looks like

bw w/no -j
957.195u 263.416s 21:11.71 95.9%        1366+1576k 11744+3435io 3031pf+0w
bw w/-j2
964.685u 294.253s 32:16.10 65.0%        1374+1570k 15335+3920io 1622pf+0w
bw w/-j4
966.085u 298.956s 26:38.43 79.1%        1337+1536k 29343+3511io 3090pf+0w
bw w/-j6
964.058u 298.116s 25:59.72 80.9%        1338+1538k 12173+3209io 2711pf+0w

The no -j was the fastest by almost 4 minutes. I have a build script 
and I say "time mkworld" and then cat >> times.log and copy the time 
output line.

Kent

-- 
Kent Stewart
Richland, WA

mailto:kbstew99@hotmail.com
http://users.owt.com/kstewart/index.html
FreeBSD News http://daily.daemonnews.org/


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3C4DF975.8050308>