[754] in java-interest
Re: trig precision problems?
daemon@ATHENA.MIT.EDU (Richard Tuck)
Tue Jul 18 20:59:02 1995
Date: Tue, 18 Jul 1995 13:28:50 +0800
From: rt@scndprsn.Eng.Sun.COM (Richard Tuck)
To: java-interest@java.Eng.Sun.COM, misfeldt@mda.ca
From: misfeldt@mda.ca (Trevor Misfeldt)
To: java-interest@java.Eng.Sun.COM
Subject: trig precision problems?
In porting C++ programs to java, I came across a
problem in the accuracy of the java trigonmetric
functions. As an example, I use sin() here.
double x = 0.5;
double y = Math.sin( 0.5 );
System.out.println( "y: " + y );
double y2 = y - 0.4794;
Your problem is here ^^^^^^
System.out.println( "y2: " + y2 );
The above code fragment produces:
y: 0.479426
y2: 2.55297e-05
which I have taken to be:
0.4794255297
BTW, the y2 variable is introduced here because I have not
figured out how to get better precision in the output of a
double.
Now, an analagous C++ program (running on the same machine,
also with a double ) produces:
0.479425538604203
Similarly, an HP scientific calculator yields:
0.479425538604
A calculator tool produces:
0.479425538604
Although I have agreement to 7 significant digits, I need
much better. I have ported a function which produces
results differing by up to 50% from the C++ implementation
because of the trig differences.
Questions
=========
Is this a display problem or a processing problem?
Can anyone else verify this?
Why does the Java sin function for double agree to only 7
significant digits with a C program when the float results
agree to 8 digits?
The sin function is fine, but the float constant is
only single precision!
Is this discrepancy known and will it be fixed in an upcoming
release or am I stuck with this accuracy?
Cheers,
Trevor Misfeldt
Consider this program:
class tsin {
public static void main( String xxxx[] ) {
double x = 0.5;
double y = Math.sin( 0.5 );
System.out.println( "y: " + y );
double y2 = y - 0.4794;
System.out.println( "y2: " + y2 );
double y3 = y - 0.4794d;
System.out.println( "y3: " + y3 );
}
}
It produces this output:
y: 0.479426
y2: 2.55297e-05
y3: 2.55386e-05
Notice how y3 differs from y2. Notice how 0.4794d differs
from 0.4794
Java float constants are single-precision by default. To make
them double, append type suffix d. The sin() function is the same.
rt
-
Note to Sun employees: this is an EXTERNAL mailing list!
Info: send 'help' to java-interest-request@java.sun.com
