A digression around VA's SRC #012b

[J-F Garnier]

(11-25-2022 02:21 PM)Werner Wrote:  Why not, in FNP:

X2 = X*X

and in the loop:

B(L)=B(L-1)*X2

should be faster than 5000 exponentials?

Yes, faster, but may introduce accumulated rounding errors, that are difficult to analyse in the final result. I chose the safest way.

J-F