I've recovered the reason. It is a feature of itensor. I'm using kdelta tensor
inside definitions of some tensors and procedures. So I had to know
I have to write KDELTA in upper case always. Do not know why it is so because
in console maxima does everything for me
(C3) display2d:false;
(D3) FALSE
(C4) show(kdelta([i,j],[l,k]))$
(D4) KDELTA[i]^l*KDELTA[j]^k-KDELTA[i]^k*KDELTA[j]^l
But not the same when kdelta is invoked inside some procedure.
So, the unevaluated kdelta made tensor huge. The situation was corrected
after replacing kdelta -> KDELTA everywhere in my procedures.
rgds,
Valery
On Tuesday 14 October 2003 18:50, Richard Fateman wrote:
RF> I would be very surprised if bothcases:false changed
RF> any speed of computation, since it is supposed to affect
RF> input only. If you are specific about benchmarks (which
RF> lisp?) it would be better, but I would look for
RF> other differences. Were they both computed in a
RF> fresh maxima?
RF>
RF> What does factor !! mean?
RF> is that a factor of 11 ?
RF> RJF