Timing, profiling...



-----Robert Dodier wrote: -----

> After glancing at src/asum.lisp I think the
> problem with sum(a[i],i,1,n) is that the partial sum is
> resimplified after each term is added.

Another fact:

   l : makelist(x[i],i,1,n)$
   apply("+",l)$

calls 'great' (n^2 - n) / 2 times. So 'simplus' could
be better, but it isn't terrible. My focus on 'simplus'
was wrong.

Barton