Date: Tue, 11 Aug 2015 15:29:54 +0200 From: Willem Jan Withagen <wjw@digiware.nl> To: Jordan Hubbard <jordanhubbard@me.com>, Dieter BSD <dieterbsd@gmail.com> Cc: freebsd-hackers@freebsd.org Subject: Re: Sparc64 support Message-ID: <55C9F8D2.5030704@digiware.nl> In-Reply-To: <16D597AE-613F-431F-8F56-30A8908F1913@me.com> References: <CAA3ZYrA0YAGtGHhcv1MJDGAdNBU31_dJ6Y0Fihpce91La=YVZQ@mail.gmail.com> <16D597AE-613F-431F-8F56-30A8908F1913@me.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 11-8-2015 10:11, Jordan Hubbard wrote: >> I see a lot of hatred towards the less popular, "weird" arches. >> Having a variety of arches has a *lot* of value. > > I’m sorry, but on top of the points above where I think you’re fairly > far off-base, you could not be more wrong here either. Even NetBSD, > which has long had the motto of “of course it runs NetBSD!” (as if > the answer was so obvious as to be unworth the question), has been > retiring architectures left and right because there is no such thing > as “free” in software. Everything has a cost in time, in complexity, > in maintenance headaches, etc. You absolutely MUST weigh the > relative value of each and every platform (or HW device) you support > and be willing to ruthlessly cull the old and the weak or before > long, your software will be a collection of burdensome conditionals > and weird constructs that no one even understands the purpose of > anymore, but “they were necessary for something, at some point” so no > one dares remove them, either. > > Just ask the OpenSSL project how heavy the burden of history can be > (and look at how many lines of code LibreSSL has ripped out, often > with great glee) and then ask yourself again if your definition of > “value” is truly aligned with the converse reality we objectively > know to be true Well it starts with the fun job of writing compiler-backends where the code is generated... Let alone that the backend writes optimized machine-code. Newer CPUs allow for combinations of instructions never considered for which new algorithms must be designed to actual be able to use them efficiently. Then it gets to the OS and the platform itself, where ARM is a real nice example.. You can call it ARM, but just only the CPU has 4 modes, let alone that there are various versions with different instruction. sets. Then go the the VM and other system architecture variations and you understand why the development of FreeBSD on ARM still has lots of very tricky changes to accommodate for. Last but not least are the devices that come with new platforms. Sure they look a lot like the ones already done, especially if they are in the PCI family. But still fine details need to be tinkered to get devices to work (flawless). And then once you have mastered that all, try to retrofit something like ZFS.... No, over the years I've always been happy that FreeBSD was deliberately careful in selecting their platforms. Because there is always NetBSD as close alternative. I install BSD stats on most of the boxed I operate, but obviously not many people do... And even less SPARC people do, so it seems. Goto http://www.bsdstats.org/bt/cpus.html to get a hang of where FreeBSD is running.... Short version: Out of 11.000 submissions: (SUNW,UltraSPARC-IIe @ 500 MHz) 5 Microsystems UltraSparc-IIe 12 Microsystems UltraSparc-IIIi 2 Microsystems UltraSparc-IIi 1 But no ARM at all, so that sort of make the numbers above fall in the range: lies, damn lies, statistics. :) --WjW
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?55C9F8D2.5030704>