Changes by: Yegor N. Bryukhov (ybryukhov at gc.cuny.edu)
Date: 2004-05-21 13:16:08 -0700 (Fri, 21 May 2004)
Revision: 5800
Log message:
Made a benchmarking:
arithT(+normalizeC) vs arithT+itt_mpoly2.standardizeT
on itt_int_arith/test* arithT+normalizeC is 12 times faster.
Partly because of lot's of wf-subgoals.
(There was an impression that a reflection based normalization
is not supposed to make a lot of wf-subgoals.)
Partly because of inefficient strategy of computation I guess.
So I have to figure out what's going on. First, I'm going to strip out
itt_mpoly2 computations into a separate resource (it's in reduce now)