Re: Same .m file: different results with different versions of Octave

[Judd Storrs]

On Mon, Apr 19, 2010 at 4:36 PM, Judd Storrs <address@hidden> wrote:
> GSL also computes sinh(2x) as sinh(x)*cosh(x) but I'm not sure if that
> results in much of an improvement in accuracy. I sort of feel like if
> it did, sinh would should use that technique internally but I'm not certain.

Thinking about this more, GSL uses this as part of a larger
expression. It's probably there to improve precision of extended
precision intermediates.


--judd