[Axiom-developer] performance issues solved

[Renaud . Rioboo]

Dear Axiom lovers,

The issue I had with the version of my Real Closure Package is solved. I was
not using the same sources for the different versions.

The issue came because version 1.1 was released too early with Axiom 2.3 and
still contained problems. When I fixed a bug in January, I fixed it from the
NAG version and not from my version which contains more code.

I have sent Tim the new version. Here is the example I used. It is due to John
Abott.

Ran := RECLOS(FRAC INT)
)se me ti on
a0 := sqrt(2,3)$Ran
a1 := sqrt(-(399063623035 - 316737007504*a0),3)
a2 := sqrt(-(5539 - 92526953*a1),3)
a3 := sqrt(-(5739 - 9473525*a2),3)
a4 := sqrt(-(1 - 1283*a3),3)
a5 := sqrt(-(4579804 - 191560867*a4),3)
a6 := sqrt(-(49400 - 220877349*a5),3)

------------------computation of a5 with Axiom 2.3------------------
                                           Type: RealClosure Fraction Integer
                         Time: 34.60 (EV) + 0.01 (OT) + 1.12 (GC) = 35.73 sec
------------------computation of a5 with Axiom 2.3------------------

and 

------------------computation of a5 with Open Axiom------------------
                                           Type: RealClosure Fraction Integer
                                     Time: 35.60 (EV) + 7.96 (GC) = 43.56 sec
------------------computation of a5 with Open Axiom------------------

Thanks to all of you for your help.

                Renaud