From owner-cvs-all Thu May 10 0:47:36 2001 Delivered-To: cvs-all@freebsd.org Received: from mailman.zeta.org.au (mailman.zeta.org.au [203.26.10.16]) by hub.freebsd.org (Postfix) with ESMTP id 738B337B423; Thu, 10 May 2001 00:47:31 -0700 (PDT) (envelope-from bde@zeta.org.au) Received: from bde.zeta.org.au (bde.zeta.org.au [203.2.228.102]) by mailman.zeta.org.au (8.9.3/8.8.7) with ESMTP id RAA16989; Thu, 10 May 2001 17:47:18 +1000 Date: Thu, 10 May 2001 17:45:47 +1000 (EST) From: Bruce Evans X-Sender: bde@besplex.bde.org To: Warner Losh Cc: Mikhail Teterin , eric@FreeBSD.org, mi@aldan.algebra.com, knu@iDaemons.org, will@physics.purdue.edu, sobomax@FreeBSD.org, kris@obsecurity.org, kris@FreeBSD.org, cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: port policies In-Reply-To: <200105091729.f49HTab38677@harmony.village.org> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-cvs-all@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG On Wed, 9 May 2001, Warner Losh wrote: > In message Bruce Evans writes: > : -j2 on a single-CPU machine with 2 disks (src on one, obj and root on the > : other) was a little slower and consumed more resources for makeworld > : the last time I checked. It might be faster because the disk(s) are too > : slow relative to the CPU and/or the VMIO cache is too small. It's hard > : to see how it could consume less resources. > > I think that people are getting confused because for a kernel build -j > 2 or -j 3 can be up to 10-15% faster than no -j flags at all. They > figure this translates well to the rest of the system. I did a quick but not very well controlled test involving "make cleandir; make obj; make depend; time make" in perl, but the result was so unfavourable for -j2 that I didn't believe it. The perl build is less gcc-bound than most of buildworld, so I would have expected it to favour -j2 more than buildworld but not be very different from makeworld. I also wouldn't expect -j2 to speed up building kernels (at least without modules) on uniprocessors. Here the build is 98% cpu-bound after the first run (all the source files fit in the disk cache). Bruce To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message