Date: Mon, 13 Dec 2004 13:00:30 -0600 From: Kirk Strauser <kirk@strauser.com> To: freebsd-alpha@freebsd.org Subject: Re: Safe (but quick) GCC settings on a PC64 with 5.3? Message-ID: <200412131300.34382.kirk@strauser.com> In-Reply-To: <83229A4B-4D2E-11D9-9C15-000D93C47836@xcllnt.net> References: <200412131130.59807.kirk@strauser.com> <83229A4B-4D2E-11D9-9C15-000D93C47836@xcllnt.net>
index | next in thread | previous in thread | raw e-mail
[-- Attachment #1 --]
On Monday 13 December 2004 11:43, Marcel Moolenaar wrote:
> I would start by raising the optimization level to 2 (as in -O2). This
> is our default now in -CURRENT and -STABLE and -CURRENT aren't really that
> far apart that it's not worth a try. You may run into an occasional
> compile breakage though, but I don't expect you'll see any instability.
Thanks for the tip. I wasn't aware that the situation had improved. At any
rate, I'm going to run a set of benchmarks (benchmarks/{bytebench, lmbench,
stream, scimarch2c}) at the various optimization levels and see if I can
find a noticeable difference.
Out of curiosity, on my Athlon systems I can set "CPUTYPE=athlon-tbird"
in /etc/make.conf and it will automatically add "-march=athlon-tbird" to my
system and ports builds. On my Alpha, though, it only adds "-mcpu=ev45"
and not the expected "-march=ev45". Is this general to all Alphas, or is
it because my ev45 is the bottom of the barrel instruction-wise so there's
no instruction set optimization to be had?
--
Kirk Strauser
[-- Attachment #2 --]
-----BEGIN PGP SIGNATURE-----
iD8DBQBBvebS5sRg+Y0CpvERAhrsAJ9cFMkWH1gyFxmEolxwflHIvgIPawCgmL78
x8PtA+tS4r7qToG8600ksNs=
=pSCc
-----END PGP SIGNATURE-----
help
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200412131300.34382.kirk>