[PPL-devel] [GIT] ppl/ppl(sparse_matrices): Fixed a performance regression caused by a bug in the computation of

[Enea Zaffanella]

Module: ppl/ppl
Branch: sparse_matrices
Commit: 95219a9f931d8b868975050d10326bde214dda46
URL:    http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=95219a9f931d8b868975050d10326bde214dda46

Author: Enea Zaffanella <zaffanella at cs.unipr.it>
Date:   Mon Jan  2 10:17:15 2012 +0100

Fixed a performance regression caused by a bug in the computation of
the number of subsumed rows.

---

 src/Polyhedron_conversion.templates.hh |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/src/Polyhedron_conversion.templates.hh b/src/Polyhedron_conversion.templates.hh
index e850214..90f0ce5 100644
--- a/src/Polyhedron_conversion.templates.hh
+++ b/src/Polyhedron_conversion.templates.hh
@@ -700,7 +700,7 @@ Polyhedron::conversion(Source_Linear_System& source,
 	      // NOTE: this number has to be less than `k' because
 	      // we are treating the `k'-th constraint.
 	      const dimension_type num_common_satur
-                = k - new_satrow.count_ones();
+                = k - redundant_source_rows.size() - new_satrow.count_ones();
 
 	      // Even before actually creating the new ray as a
 	      // positive combination of `dest_rows[i]' and `dest_rows[j]',