[Fwd: QR-Factorization in gsl (fwd)]

[Gerard Jungman <jungman at lanl dot gov>]

-- 
Gerard Jungman <jungman@lanl.gov>
Los Alamos National Laboratory

--- Begin Message ---

• From: Joerg Wensch <wensch at informatik dot uni-halle dot de>
• To: jungman at lanl dot gov
• Date: Fri, 19 Mar 2004 16:43:30 +0100 (CET)
• Subject: QR-Factorization in gsl (fwd)

Dear Collegue,

we were using the QR-decomposition written by you
for the gsl. We recogniced that the row-oriented
format decreases performance considerably for larger
matrices. We had a 10000 by 160 matrix were
the composition took very long (about 45 sec.).

It is clear to that it is a lot of extra work
to allow both formats (row+col major)
in the matrix class. I propose a simple workaroud:
To add besides an QR-decomposition an LQ-decomposition too.

I have used an LQ-decomposition for the large matrix
(now in 160 times 10000 format) which gave a speedup of
13 on our architecture. I have attached the source-code
for the non-pivotized version (which I had obtained
by slight modifications of your source code -- I have left your
copy-right notice on the top of the file)
May be you want to accept this as a contribution to gsl.
I am also willing to contribute transposed versions for
pivotized algorithms when I find a time slot for doing this.

Regards,
Joerg.

P.S. I resend this because I forgot to attach the file ...
------------------------------------
Dr. Joerg Wensch
MLU Halle
FB Mathematik und Informatik
Institut fuer Informatik
06099 Halle
Germany
------------------------------------

Attachment: gsl_LQ.cc
Description: Text document

--- End Message ---