OON: Re: BLITZ, TNT and GNUSSL

From: Dave Steffen (steffend@glitch.physics.colostate.edu)
Date: Tue Mar 03 1998 - 12:19:58 EST


Evgenii Rudnyi wrote:

> I have got to choose matrix C++ library for my work. From what I
> have seen, BLITZ, TNT and GNUSSL are the leaders among numerous
> matrix classes available on Internet.
>
> It happened that I have started with GNUSSL. Certainly, nothing is
> perfict but I was able to compile the tests, I like the idea behind
> it, I seem to be able to understand the code. Still, I am not sure
> which is the best choice.
>
> If someone has experience with these packages and can share thoughts
> on comparison between them, I would greatly appreciate it.

  For what it's worth:

  I've been using TNT for a while now, and I like it. It's a fairly
  straightforward library (which means it's simple enough for me to
  hack ;-) but it's also fairly complete. It's also got the potential
  to evolve into something with a _lot_ of fancy optimizations, in
  particular it'll eventually be seamlessly integrated with the BLAS,
  and LAPACK libraries, which are _really_ quick.

  It's not too hard to integrate TNT and LAPACk routines _now_, if you
  need to. My numerical code is based on doing exactly this, and it's
  relatively easy to set up.

  BLITZ, on the other hand, is very complicated - I don't understand
  how it works, even remotely. It uses fancy template stuff to do
  things like evaluate the sum of three vectors in one loop, instead
  of two. It's a nice thought (if your compiler can handle it), but
  it's too complicated for me to hack easily, and it's not complete
  yet - the vectors are done, but I don't think they've gotten around
  to the matrices yet.

  I'd hate to try to figure out how to call LAPACK routines with BLITZ
  objects - it's probably possible, and maybe it's easy, but I don't
  know for sure.

  I don't know much about GNUSSL, so I can't comment there.

  Just my $.02 (and possibly overpriced! ;-)

--------------------------------------------------------------------------
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@lamar.colostate.edu 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



This archive was generated by hypermail 2b29 : Wed Feb 20 2002 - 03:20:06 EST