Boost logo

Boost :

From: Dave Steffen (steffend_at_[hidden])
Date: 2001-03-14 13:16:28

lums_at_[hidden] writes:
> --- In boost_at_y..., "David Abrahams" <abrahams_at_m...> wrote:
> > Is the idea of building in the ability to dispatch to BLAS based
> > on any real measurements, or just the assumption that BLAS will
> > be faster than anything we can reasonably do in C++? Jeremy's
> > paper on MTL seems to indicate that involving BLAS is
> > unneccessary. Overgeneralizing is one sure way to kill a
> > software project.
> I think there is one reason one might want to dispatch to the BLAS
> -- to escape to a performance layer while we await other compilers
> to catch up with KCC (this was suggested elsewhere in another post
> by another individual -- neither of which I can recall or locate
> right now).
> In other words, this would only be temporary.

 Right. Plus, if someone's done a really good job of generating an
 optimized BLAS library for your machine --- one that _still_
 outperforms our fancy C++ library --- you could still use it.

 Of course, it might be more work to build in BLAS dispatch than it
 would be to do the C++ right, in which case never mind. ;-) If we
 end up using MTL (the conversation seems to be drifting that way),
 then it's probably a moot point. ;-)

Dave Steffen Wave after wave will flow with the tide
Dept. of Physics And bury the world as it does
Colorado State University Tide after tide will flow and recede
steffend_at_[hidden] Leaving life to go on as it was...
                                                        - Peart / RUSH
"The reason that our people suffer in this way....
is that our ancestors failed to rule wisely". -General Choi, Hong Hi

Boost list run by bdawes at, gregod at, cpdaniel at, john at