Optimizing sparse matrix operation

Say I am trying to compile a language which has sparse matrix and vector
types when I have an expression such as:


      for(...){
              ...
              y = A*B*x;
              ...
              }


where A and B are sparse matrices and x and y vectors. I suppose A and B
are constant over the loop, so I could evaluate y as A*(B*(x)) or
precompute C=A*B before the loop and compute y as C*x. Depending on the
sparcity structure of A and B, C will or not be sparse so this
optimization can or not be acceptable. Making this decision might also be
costly.


So my question is: has anybody worked on this type on problem and if so
where can I find some references.


James Litsios
--
James Litsios Phone: +41 1/256 49 50
Integrated Systems Laboratory Fax: +41 1/252 09 94
ETH Zurich E-Mail: litsios@iis.ethz.ch
CH-8092 Zurich, Switzerland
--

Return to the comp.compilers page.
Search the comp.compilers archives again.