[GiNaC-devel] Restore the speed of polynomial operations.

[Alexei Sheplyakov]

Hello!

These patches

commit 0633ed8082961673eedc092689e06fa39d6bc322
Author: Jens Vollinga <vollinga at thep.physik.uni-mainz.de>
Date:   Thu May 19 18:10:40 2005 +0000

    Fixed bug in expanding expressions containing dummy indices. [V.Kisil]

(see <http://www.ginac.de/pipermail/ginac-cvs/2005-May/001586.html>)

commit 6b1b11264fe4872132fef8cb8b22cb983ba01736
Author: Chris Dams <Chris.Dams at mi.infn.it>
Date:   Thu Jul 13 20:02:38 2006 +0000

    Needed to call get_all_dummy_indices_safely more often.

(see http://www.ginac.de/pipermail/ginac-cvs/2006-July/001780.html)

commit 4cf53676d7823ec3c55a90df7723f7f1492a95b2
Author: Chris Dams <Chris.Dams at mi.infn.it>
Date:   Mon Jul 24 16:47:43 2006 +0000

    Improved dummy index renaming.

(see http://www.ginac.de/pipermail/ginac-cvs/2006-July/001782.html)

made polynomial operations (in particular expand()) substantially slower.
For example, Fateman's polynomial expanding benchark degraded from 32sec.
to 80sec. My own code which uses multivariate GCD (which in turn uses 
.expand() extensively) become 4x slower.

I'll post 4 patches which address the issue in reply to this message.
Comments, questions, suggestions, etc. are wellcome.

Best regards,
  Alexei

-- 
All science is either physics or stamp collecting.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 827 bytes
Desc: Digital signature
Url : http://www.cebix.net/pipermail/ginac-devel/attachments/20070917/2ebd6fac/attachment.pgp