Date: Tue, 10 Feb 2009 17:07:03 +0100 From: Nick Hibma <nick@van-laarhoven.org> To: svn-src-all@freebsd.org Subject: Re: svn commit: r188439 - head/release Message-ID: <200902101707.03977.nick@van-laarhoven.org> In-Reply-To: <9bbcef730902100753p484d633j29829019064e71bf@mail.gmail.com> References: <200902101511.n1AFBQXx096922@svn.freebsd.org> <1234279584.65150.13.camel@bauer.cse.buffalo.edu> <9bbcef730902100753p484d633j29829019064e71bf@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
> 2009/2/10 Ken Smith <kensmith@cse.buffalo.edu>: > > I wouldn't mind if it got turned on but that's just the first of two > > questions. The second question is what to set it to. What's best for > > that depends on how many CPUs you've got and/or how much of your > > machine you want the build to consume (you might want to be doing other > > stuff on the machine while the build is running...). IMHO it might be > > best to just leave it as-is and let anyone interested in getting better > > performance set it to what's appropriate for their machine but if > > others feel differently I won't object. > > I was think more about the semi-supported state that the -jX > traditionally has - it's usually something like "yes, it's there and > probably everyone uses it but we won't even suggest it actually > works". Having it on in the official release process will signal > official support to users and encourage them to use it. With the advent of multiple cores it is well worth doing at least a -j`sysctl -n kern.smp.cpus` (assuming that value is always available, even in the non-SMP case) so the parallel build gets exercised. Nick
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200902101707.03977.nick>