[boost] uBLAS vs MTL for sparse matrices

From: Paul C. Leopardi (leopardi_at_[hidden])
Date: 2002-05-02 06:22:49

Hi all,
I am currently developing GluCat, a generic library of universal Clifford 
algebra templates - see http://glucat.sf.net
GluCat 0.0.6 uses MTL 2.1.2-20. 

GluCat uses MTL because of its support for sparse matrices.
GluCat often needs to manipulate matrices which have the same sparsity 
pattern as permutation matrices, eg. 1024 x 1024 with 1024 non-zeros.
To do this, GluCat routines use MTL iterators which provide iteration over 
the non-zero entries of a matrix. This means GluCat can iterate over a sparse 
n x n matrix in O(n) time rather than O(n^2).

I am thinking of making GluCat as Boost-compatible as possible, and of 
improving its readability. This would eventually involve operators, concepts, 
etc. and may involve a switch from MTL to uBLAS, especially if MTL 3 is not 
being actively developed.

I have read that uBLAS now supports sparse matrices. Does the sparse matrix 
data structure used in uBLAS povide the same sort of iterators as MTL?
Is it possible to iterate over an n x n permutation matrix in O(n) time? How?
Thanks