Date: Mon, 14 Nov 2005 19:21:32 -0500 From: Kris Kennaway <kris@obsecurity.org> To: Chris <chris@childeric.freeserve.co.uk> Cc: freebsd-questions@freebsd.org Subject: Re: Makefile and Kernel Optimizations in 6.0 Message-ID: <20051115002132.GB67157@xor.obsecurity.org> In-Reply-To: <43792010.9010802@childeric.freeserve.co.uk> References: <20051114102308.7515.GERARD@seibercom.net> <20051114202300.GB98749@xor.obsecurity.org> <43792010.9010802@childeric.freeserve.co.uk>
next in thread | previous in thread | raw e-mail | index | archive | help
--p4qYPpj5QlsIQJ0K Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Mon, Nov 14, 2005 at 11:38:56PM +0000, Chris wrote: > Kris Kennaway wrote: > >P.S. -funroll-loops and -ffast-math are commonly used by people who > >don't understand what they mean but like the thrill it gives them to > >compile their application with secret optimizations (e.g. the word > >"fast" makes them feel excited). However, the former is often not > >actually an optimization (which is why it's not on by default), and > >the latter can cause incorrect computations (but I think it's > >completely irrelevant in the kernel which does not do FP math). >=20 > Just out of curiosity, what would people who know what they are doing=20 > use them for? -funroll-loops sounds like an amusement park :P They would carefully test to see whether it helped their specific CPU-intensive application, and only use it if so. Kris --p4qYPpj5QlsIQJ0K Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (FreeBSD) iD8DBQFDeSoLWry0BWjoQKURAkI7AKDuQoCEJyDL7C1ivx5PspPygCkBawCeLUAf tbBXdX3g1mzyJ4QvlHjZUI0= =IGxl -----END PGP SIGNATURE----- --p4qYPpj5QlsIQJ0K--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20051115002132.GB67157>