Trig functions

[Jon Rea]

Im looking to do some heavy geomentry based number crunching involving a lot
of sin and cos calculations.

Performing these operations takes up a lot of computer time, we have
calculated ~100 clock cycles per calculation.

I have come accross approximations such as the "Taylor Series" for working
out the values of these functions. What precision are sin and cos calculated
by the computer? does anyone know how the computer calculates sin and cos ?
(i read a while back there is an efficient binary method, but i lost the
site and the reference to the name of the method). I assume that the
processors tend to have special high speen pipelines for this sort of thing
too ?

If i only want sin and cos calculated to float precision, are there actually
better ways of calculating sin and cos for number crunching applications.

Many thanks for your time
Jon Rea

[Ignacio Machin]

Hi Jon,

Have you think in using a C dll for this? I don't know if you will find a
good and proven math component written in C#, if you application is very
trig. intensive I would strongly advise that you use one library done in C
this this task.
Do a google search and you probably will find several libs like that.

Hope this help,

--
Ignacio Machin,
ignacio.machin AT dot.state.fl.us
Florida Department Of Transportation

"Jon Rea" <jr****@bris.ac.uk> wrote in message
news:XP*******************@news-binary.blueyonder.co.uk...

Im looking to do some heavy geomentry based number crunching involving a lot of sin and cos calculations.

Performing these operations takes up a lot of computer time, we have
calculated ~100 clock cycles per calculation.

I have come accross approximations such as the "Taylor Series" for working
out the values of these functions. What precision are sin and cos calculated by the computer? does anyone know how the computer calculates sin and cos ? (i read a while back there is an efficient binary method, but i lost the
site and the reference to the name of the method). I assume that the
processors tend to have special high speen pipelines for this sort of thing too ?

If i only want sin and cos calculated to float precision, are there actually better ways of calculating sin and cos for number crunching applications.

Many thanks for your time
Jon Rea

[Joubert]

Hi Jon,

try out http://www.lundin.info/mathparser.asp

Regards

Joubert
"Jon Rea" <jr****@bris.ac.uk> wrote in message
news:XP*******************@news-binary.blueyonder.co.uk...

Im looking to do some heavy geomentry based number crunching involving a lot of sin and cos calculations.

Performing these operations takes up a lot of computer time, we have
calculated ~100 clock cycles per calculation.

I have come accross approximations such as the "Taylor Series" for working
out the values of these functions. What precision are sin and cos calculated by the computer? does anyone know how the computer calculates sin and cos ? (i read a while back there is an efficient binary method, but i lost the
site and the reference to the name of the method). I assume that the
processors tend to have special high speen pipelines for this sort of thing too ?

If i only want sin and cos calculated to float precision, are there actually better ways of calculating sin and cos for number crunching applications.

Many thanks for your time
Jon Rea