Date: Thu, 16 Aug 2007 12:13:21 -0400 From: Mikhail Teterin <mi+mill@aldan.algebra.com> To: "Jeremy Messenger" <mezz7@cox.net> Cc: ahze@freebsd.org, multimedia@freebsd.org Subject: `make -jN' (Re: minor improvement to devel/liboil) Message-ID: <200708161213.21594.mi%2Bmill@aldan.algebra.com> In-Reply-To: <op.tw5s1pjm9aq2h7@mezz.mezzweb.com> References: <200708151254.l7FCs2BH098003@aldan.algebra.com> <op.tw5s1pjm9aq2h7@mezz.mezzweb.com>
next in thread | previous in thread | raw e-mail | index | archive | help
четвер 16 серпень 2007 11:52 до, Jeremy Messenger Ви написали:
> The -jN is well known for cause problems on some systems and it
> requires to do a lot of tests on different arch, 4BSD and ULE schedulers
> on both -STABLE and -CURRENT on HTT, multi-core, and multi-CPU (physical
> CPU) systems. marcus has a problem with graphics/jasper until he removed
> this same line to solve his problem on his dual-core amd64 box.
Khmm, the only problems I'm aware off are poorly-written Makefiles -- and
jasper's was Ok. Are you alluding to a bug in a scheduler?
Actually, a better change is to amend the ALL_TARGET instead of MAKE_ARGS:
ALL_TARGET= -j`${SYSCTL} -n hw.ncpu`
This allows the port to _build_ faster, but still _install_ serially.
Install-targets create directories and that does not parallelize well in many
autotools-created Makefiles.
Maybe, that was Marcus' problem?
-mi
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200708161213.21594.mi%2Bmill>