From owner-freebsd-questions Fri Mar 24 19:54:25 2000 Delivered-To: freebsd-questions@freebsd.org Received: from mail.rdc1.ne.home.com (ha1.rdc1.ne.home.com [24.2.4.66]) by hub.freebsd.org (Postfix) with ESMTP id F223437B7D7; Fri, 24 Mar 2000 19:54:11 -0800 (PST) (envelope-from jgowdy@home.com) Received: from cx443070a ([24.4.93.90]) by mail.rdc1.ne.home.com (InterMail v4.01.01.00 201-229-111) with SMTP id <20000325035402.HHIR14878.mail.rdc1.ne.home.com@cx443070a>; Fri, 24 Mar 2000 19:54:02 -0800 Message-ID: <000b01bf960f$0693f960$0100000a@vista1.sdca.home.com> From: "Jeremiah Gowdy" To: "Doug Barton" , "Christian Weisgerber" Cc: , References: Subject: Re: PR #17479: parallel building of ports Date: Fri, 24 Mar 2000 20:03:05 -0800 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.00.2919.6700 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6600 Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG > > I received absolutely *no* reaction in response to PR ports/17479 > > which suggests a minimal change to allow reliable parallel building > > (i.e. make -jX) of ports. > > > > I can't imagine that I'm the only person with an SMP box who cares > > about this. > > You will probably get more of a reaction by goading the folks in > -ports. :) Hey, I'm with you all the way. It just plain sucks when you want to take full advantage of your dual processor setups (I have two of them) and you get 3/4 of the way through the build (which is hauling serious butt, especially with SMP *and* SOFTUPDATES to prevent massive harddrive head strokes) then suddenly have your build fail because something was built out of order. They could at the LEAST have an option in the Makefile that said whether it was safe to do -jX. Then you could just set your setup to do -jX by default, and any Makefiles that don't allow -jX (because it doesn't work for that build) would override the option. That would be the MINIMUM they could do. I honestly agree they should totally make an effort to make their Makefiles -jX compatible. With SMP becoming so popular as you can get a Dual Celeron board for $120 and a pair of Celeries real cheap, more and more people are going to be using SMP setups. If people can't take full advantage of their SMP, because of issues like kernel threads, or incompatible Makefiles, they're probably going to look for an operating system that has more flexability with SMP. I have a dual PII 333, which runs like a dream when programs take full advantage of it. Apache and Samba spawn processes, which can and do use either CPU. I can (usually) buildworld and installworld with -j 4. Installworld with -j 4 and Softupdates is absolutely amazing. SMP can be a dream come true for people who could never afford to buy a motherboard/cpu in the 600-700mhz range, but can easily afford a cheap dual SMP setup. I say at least have an option or marking that specifies -jX compatibility, and then any of the port authors who care enough can try to make some optimization/change to allow it when they're ready. Better yet, if we got a system working, that Makefiles specify if they're multijob compatible, then make itself could detect your SMP setup, and maybe the load conditions (and an override option), and decide to do -jX on it's own. There's alot more territory to be explored here with this, and all I can say is, I hate to compile something on a Pentium II 333, when I have another Pentium II 333 sitting there idle. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message