This is the mail archive of the libc-alpha@sources.redhat.com mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Implementatoinof Sin and Cos


On Tue, 2003-07-15 at 08:06, Stefan Jones wrote:
> Taylor Series converges very slowly in my opinion. Such functions are
> normally done via a small table lookup and some sort of converging
> iteration ( at least the floating point division unit I worked on does).
> 
> Most floating point calculations are done in hardware in the FPU, I
> think trigonometric function are one of them.

Replying to one's post is not good ...

It seems I was right, after a quick source browse.

But for archs without a good fpu use 

sysdeps/ieee754/dbl-64 ( I think, or just give errors )

for there FPU ops. Most of the functions use, as I said, a table lookup
and a converging iteration.

Stefan


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]