Date: Mon, 22 Jun 2020 17:09:49 +0200 From: Ralf Mardorf <ralf-mardorf@riseup.net> To: freebsd-questions@freebsd.org Subject: Re: Ports are faster than packages? Message-ID: <20200622170949.293074ec@archlinux> In-Reply-To: <20200622145225.3799de459aebc076a8dbdea6@sohara.org> References: <1592833182.722746222@f403.i.mail.ru> <20200622145225.3799de459aebc076a8dbdea6@sohara.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, 22 Jun 2020 14:52:25 +0100, Steve O'Hara-Smith wrote: >On Mon, 22 Jun 2020 16:39:42 +0300 >nikitastepanov113@yandex.ru wrote: > >> Ports are faster than packages? > >Unless you have a very very fast build server and a slow >connection to the world packages will generally be faster to install >and maintain. The bigger the package the more true this is - things >like Gnome, KDE, Firefox, LibreOffice take a lot of time and resources >to build these days. Even the fastest build machine would suffer from a slow Internet connection, too ;). Assuming the request should be related to the performance of the software, then software build on your machine, build to fit to your hardware, could be faster, than software provided by a package, that needs to fit to e.g. processors with different instruction sets. Imagine software build to apply to SSE vs software build with SSE"+". Imagine performance must fit to hard real-time requirments. If so, something build with an instruction set that fits to everybody's computer, could even provide less good hard real-time results of measurements or results in less good audio quality etc. ... However, almost all of the times it's way faster to install a package, than to build from source and in most cases optimisation doesn't really matter. Even for hard real-time ability and similar corner cases, it only sometimes might make a noticeable difference.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20200622170949.293074ec>