Abstract In this work an efficient algorithm to perform a block decomposition for large dense rectangular matrices with entries in F2 is presented. Matrices are stored as column blocks of row major matrices in order to facilitate rows operation and matrix multiplications with blocks of columns. One of the major bottlenecks of matrix decomposition is the pivoting involving both rows and column exchanges. Since row swaps are cheap and column swaps are order of magnitude slower, the number of column swaps should be reduced as much as possible. Here an algorithm that completely avoids the column permutations is presented. An asymptotically fast algorithm is obtained by combining the four Russian algorithm and the recursion with the Strassen algorithm for matrix–matrix multiplication. Moreover optimal parameters for the tuning of the algorithm are theoretically estimated and then experimentally verified. A comparison with the state of the art public domain software SAGE shows that the proposed algorithm is generally faster.

Fast matrix decomposition in F2

Bertolazzi, Enrico;Rimoldi, Anna
2014-01-01

Abstract

Abstract In this work an efficient algorithm to perform a block decomposition for large dense rectangular matrices with entries in F2 is presented. Matrices are stored as column blocks of row major matrices in order to facilitate rows operation and matrix multiplications with blocks of columns. One of the major bottlenecks of matrix decomposition is the pivoting involving both rows and column exchanges. Since row swaps are cheap and column swaps are order of magnitude slower, the number of column swaps should be reduced as much as possible. Here an algorithm that completely avoids the column permutations is presented. An asymptotically fast algorithm is obtained by combining the four Russian algorithm and the recursion with the Strassen algorithm for matrix–matrix multiplication. Moreover optimal parameters for the tuning of the algorithm are theoretically estimated and then experimentally verified. A comparison with the state of the art public domain software SAGE shows that the proposed algorithm is generally faster.
2014
Bertolazzi, Enrico; Rimoldi, Anna
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11572/34517
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 4
  • ???jsp.display-item.citation.isi??? 4
social impact