On Sun, 4 Feb 2001, Kerstin Hoef-Emden wrote:
> So you are recommending either to use only -mcpu= or to use both with
> -mcpu= first?
Just use -mcpu. It should do the right thing in all cases.
> What are the "symptoms" that a special code needs -mieee? Most packages
> which I compile, are used for molecular phylogeny. MrBayes, as the name
> says, works with Bayesian methods. I guess, that fpu-operations are
> heavily used in these kind of calculations.
Yes, it sounds like the stuff that you want to compile is definitely in
the "needs -mieee" category :-)
Honestly, if the packages are meant only for local installation (and not
upload to Debian or something similar), I would recommend compiling them
with Compaq's compilers since they schedule code for EV5 and EV6 much
better than gcc probably ever will. The compilers are not free to
redistribute (hence their exclusion from Debian), but they can be obtained
from Compaq's ftp site and only require minor tweaking to work on Debian
> Unfortunately, computing time is an important item in molecular
> phylogeny. It is a difference whether an analysis takes 12 days or 7.
Definitely use Compaq's compilers...you won't regret it :-)
> Or smaller soft which does a lot of repetitions?
> Phylogenetic methods have usually repetitive structure. Datasets are
> calculated through again and again.
Again, Compaq's compilers do VERY GOOD scheduling of loops in particular
(all loads, then operations, then stores in most cases...very efficient
> Apart from this:
> I encounter problems to report the -O-bug to firstname.lastname@example.org and
> have no idea what is going wrong. The return of my first attempt was
> understandable, I put a blank line before the Package line, but after
> correction, it still returned with error message:
> Your message didn't have a Package: line at the start (in the
> pseudo-header following the real mail header), or didn't have a
> pseudo-header at all.
> These were my starting lines:
> Package: gcc
> Version: 1:2.95.2-13
> So what is wrong about it? (Preprocessed file was attached to the
Omit the epoch (1:). I don't know why, but I think the bug system chokes
on that. Either that, or if you used a tab instead of a space between the
'Package:' and 'gcc', change it to a space. Here's a good start: