Hi, I'm working on a 2d physics engine and want to use the LAPACK routines for performance reasons.
Your library looks very interesting.
I need to be able to solve systems of linear equations. AFAIK, the best way to do this in LAPACK is to first obtain a LU-factorization using sgetrf, and then use the result of that to solve a certain vector using sgetrs.
I noticed that the source code of your library does have an implementation for sgetrf, but not for sgetrs. Why is that and how can we solve numerous vectors in a high-performance way for one matrix?
Kind regards,
Bas
Hi, I'm working on a 2d physics engine and want to use the LAPACK routines for performance reasons.
Your library looks very interesting.
I need to be able to solve systems of linear equations. AFAIK, the best way to do this in LAPACK is to first obtain a LU-factorization using sgetrf, and then use the result of that to solve a certain vector using sgetrs.
I noticed that the source code of your library does have an implementation for sgetrf, but not for sgetrs. Why is that and how can we solve numerous vectors in a high-performance way for one matrix?
Kind regards,
Bas