Changes by: Yegor N. Bryukhov (ybryukhov at gc.cuny.edu)
Date: 2004-06-09 19:55:35 -0700 (Wed, 09 Jun 2004)
Revision: 5873
Log message:
Interestingly that for this version lazy strategy (like in reduceC)
is more efficient than more eager strategy used in itt_mpoly2.
Now itt_mpoly3_bench/test* is 4 times slower than itt_int_arith/test*.
Actually the most complicated "testn" only 2 times slower but simpler
examples have obviously worse factor.
I realized that we can't expect same performance from this approach
as normalizeC has because normalizeC is combination of ml-code and
rewrites but itt_mpoly* use only rewrites for normalization.
Changes | Path |
+3 -3 | metaprl/theories/itt/itt_mpoly3.ml |