Hi all,
I'm running Singular 3-1-6 on OSX 10.8.2 with the OSX tarballs available on the site, and all is well. I'm also trying to implement some custom functions, and that's where the problems arise. I have, for example, a proc to do a Cholesky factorization on a square matrix of variable size, and having that code on a LIB is proving to be insufficient in terms of speed. The Cholesky factorization is O(n^3/3), while LU decomposition is O(2n^3/3), twice the cost of Cholesky, but LU decomp is builtin Singular and is much faster than my Cholesky proc.
So my next step is coding my procs in C++ (like the LU decomp code in the LinearAlgebra.c file) and then compiling it into Singular and use it like that, where I expect a major gain in performance.
The problem is there isn't a guide to compile Singular on OSX from the source. I've tried but I get errors in ./configure, like:
Code:
./configure: line 2194: ,: command not found
configure: error: can not find gmp
So, what I need is to know the exact procedure used to compile the OSX versions of Singular, from (I assume) the github downloadable source.
Thanks !