Cory O'Connor wrote:
Hello,
My name is Cory O'Connor and I'm a graduate Aerospace engineering
student at Cal Poly San Luis Obispo. I'm working on a grant project
doing C++ utility analysis and simulation for a spacecraft. I needed a
generalized linear equation solver due to a Least Squares problem where
there are more equations than unknowns and I'd additionally like to be
able to solve eigenvalue problems. I have read and heard from industry
contacts that LAPACK is the way to go. I note however, that LAPACK++
doesn't have eigenvalue support in it. I assume that LAPACK is the only
version of the algorithms which supports both eigenvalue problems and
singular value decomposition. Given that I'm not quite sure how to
proceed in downloading, installing and using the original LAPACK
software. Any help you could give me would be greatly appreciated.
Thanks,
Cory O'Connor
Cory,
LAPACK has routines for solving linear systems, linear least square
problems, eigenvalue problems and more. We do not support LAPACK++.
You should go with LAPACK. LAPACK is written entirely in Fortran 77.
You should be able to call LAPACK routines from a C++ program. You
will have to follow Fortran calling convention with passing of
arguments by reference. We have been working on a C interface to
LAPACK with C calling convention and passing of arguments by value
whenever possible. We have not released it yet though. We do not
envision a C++ interface, since it would not offer any significant
advantages over the C interface and using C libraries from C++ code
is common.
Check if LAPACK is not pre-installed on your machine. If it is not,
download it and go to the README file for instructions. Good luck.
Jakub
--
Jakub Kurzak -- University of Tennessee
kurzak@Domain.Removed -- (865) 974-9985
|