7+ days of dogfood
mexas at bristol.ac.uk
Tue Feb 12 11:08:08 UTC 2013
Date: Tue, 12 Feb 2013 02:48:49 -0800 (PST)
From: Jakub Lach <jakub_lach at mailplus.pl>
To: freebsd-current at freebsd.org
Hi Anton, I recognize you from your
usual struggle with ia64 :)
Hope your work/environment really have a good
justification for keeping them around...
I'm a stubborn donkey.
As lons as Marcel has not abandoned it,
I'll use it.
Regarding work/environment - the University
has become about as stale, boring, risk averse,
unimaginative as can be, with regard to
computers, computing enviroment and general IT
support attitudes. What is a poor user to do?
I make a fuss. I complain to anybody and everybody.
I'm trying to convice the director of IT and
his varioius deputies, that they must recognise
the diversity of computing needs in engineering.
Windows desktop just doesn't do it for everybody.
More to the point, I belive by using ia64,
I contribute, in a small way, to the development
of FreeBSD, e.g. sometimes ia64 code breakages
indicate poor coding, and fixing code on ia64
actually fixes it on other arches too.
With a massive help from linimon@ we managed
last time to build over 15k ports on ia64 9 and 10.
So things are not that bad.
I would actually set only -march= because it should
set mtune to the same value by default too.
Moreover, I suspect mckinley is an alias for itanium2,
as it was it's codename...
Strangely, current gcc documentation for ia64 is
missing references of march and mtune for ia64.
There's no "-march", but for "-mtune" it says:
Tune the instruction scheduling for a particular CPU, Valid values are itanium, itanium1, merced, itanium2, and mckinley.
More information about the freebsd-current