Subject: Increasing the accuracy of Gamma for double float
From: Raymond Toy
Date: Mon, 20 Oct 2008 16:17:55 -0400
Dieter Kaiser wrote:
> So, what do you think? Should we improve the accuracy and do a call to bffac
> with an preciscion of fpprec:20? Is it worth to lost a speed of 50 to gain the
> missing 3 digits in double float precision? What about the bug we have with GCL?
I think there is an expectation that float arithmetic is reasonably
fast, so making it much slower breaks that expectation. If people are
not happy with the float accuracy, they can always use bffac themselves
to get the desired accuracy.
Maybe we can use a different algorithm for real gamma values. There is
a gamma routine in slatec. Presumably it is more accurate than what we
have now.
Ray