CERFACS CERFACS Scientific Activity Report Jan. 2010 â Dec. 2011
CERFACS CERFACS Scientific Activity Report Jan. 2010 â Dec. 2011
CERFACS CERFACS Scientific Activity Report Jan. 2010 â Dec. 2011
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
4 Dense and Sparse Matrix Computations<br />
4.1 Analysis of the solution phase of a parallel multifrontal<br />
approach.<br />
P. Amestoy : ENSEEIHT-IRIT, UNIVERSITY OF TOULOUSE, France ; I. S. Duff : RUTHERFORD<br />
APPLETON LABORATORY AND <strong>CERFACS</strong>, UK and France ; A. Guermouche : LABRI, UNIV.<br />
BORDEAUX 1 / INRIA BORDEAUX SUD-OUEST, France ; Tz. Slavova : <strong>CERFACS</strong>, France<br />
In [ALG9], we study the forward and backward substitution phases of a sparse multifrontal factorization.<br />
These phases are often neglected in papers on sparse direct factorization but, in many applications, they can<br />
be the bottleneck so it is crucial to implement them efficiently. In this work, we assume that the factors have<br />
been written on disk during the factorization phase, and we discuss the design of an efficient solution phase.<br />
We will look at the issues involved when we are solving the sparse systems on parallel computers and will<br />
consider in particular their solution in a limited memory environment when out-of-core working is required.<br />
Two different approaches are presented to read data from the disk, with a discussion on the advantages and<br />
the drawbacks of each. We present some experiments on realistic test problems using an out-of-core version<br />
of a sparse multifrontal code called MUMPS (MUltifrontal Massively Parallel Solver).<br />
4.2 On computing inverse entries of a sparse matrix in an out-of-core<br />
environment.<br />
P. Amestoy : INPT(ENSEEIHT)-IRIT, UNIVERSITY OF TOULOUSE, France ; I. S. Duff : RUTHERFORD<br />
APPLETON LABORATORY AND <strong>CERFACS</strong>, UK and France ; Y. Robert : <strong>CERFACS</strong>, France ; F.-<br />
H. Rouet : INPT(ENSEEIHT)-IRIT, France ; B. Uçar : ENS LYON AND CNRS, France<br />
The inverse of an irreducible sparse matrix is structurally full, so that it is impractical to think of computing<br />
or storing it. However, there are several applications where a subset of the entries of the inverse is required.<br />
Given a factorization of the sparse matrix held in out-of-core storage, we show how to compute such a subset<br />
efficiently, by accessing only parts of the factors. When there are many inverse entries to compute, we need<br />
to guarantee that the overall computation scheme has reasonable memory requirements, while minimizing<br />
the cost of loading the factors. This leads to a partitioning problem that we prove is NP-complete. We also<br />
show that we cannot get a close approximation to the optimal solution in polynomial time. We thus need to<br />
develop heuristic algorithms, and we propose : (i) a lower bound on the cost of an optimum solution ; (ii) an<br />
exact algorithm for a particular case ; (iii) two other heuristics for a more general case ; and (iv) hypergraph<br />
partitioning models for the most general setting. We illustrate the performance of our algorithms in practice<br />
using the MUMPS software package on a set of real-life problems as well as some standard test matrices. We<br />
show that our techniques can improve the execution time by a factor of 50. For more details, see [ALG33].<br />
<strong>CERFACS</strong> ACTIVITY REPORT 11