Re: problem with cos ( pi/2 )
Re: problem with cos ( pi/2 )
- Subject: Re: problem with cos ( pi/2 )
- From: Christopher Henrich <email@hidden>
- Date: Fri, 16 Jun 2006 12:33:25 -0400
On Jun 16, 2006, at 11:07 AM, billmonk wrote:
Not sure if this is the right list...searching it hasn't turned up
any likely entries...
When calling cos( pi/2 ), I'm getting a tiny number, rather than
zero, for
cos( M_PI_2 ); // pi/2, 1.57079632679489661923 from <math.h>
I've added "-llibm" to "Other C flags", as alluded to here,
http://developer.apple.com/documentation/Darwin/Reference/ManPages/
man3/intro.3.html
Oops; if you're using C (C++, Obj-C, whatever) then I think you want
the "-lm" flag.
but am still getting 2.6484540326036093e-14 instead of 0.
My old IM:PowerPC Numeric book, and the OS X pdf here
http://developer.apple.com/documentation/Performance/Conceptual/
Mac_OSX_Numerics/Mac_OSX_Numerics.pdf
say cos( pi/2 ) is a special case (for all the reasons given in
PPCNumerics) but that an IEEE-compliant math lib should return
-0.0, with the "inexact" flag set, for cos( pi/2 ).
So it looks like I'm not actually using the IEEE libm. What am I
doing wrong?
As a data point, I notice Calculator.app in Degrees mode correctly
gives 0 for cos( 90 ), but in Radians mode gives a tiny number for
cos ( pi /2 ); a different one than I get for cos( M_PI_2 ). That's
kind of weird...
"90" is an integer, exactly representable in almost any number
format. So here's a case where the machine knows exactly what angle
you want. Therefore it can give you exactly the cosine of it.
Chris Henrich
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Xcode-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden