frexp
Split a double-precision floating-point number into a normalized fraction and an integer power of two.
Usage
var frexp = require( '@stdlib/math/base/special/frexp' );
frexp( x )
Splits a double-precision floating-point number into a normalized fraction and an integer power of two.
var out = frexp( 4.0 );
// returns [ 0.5, 3 ]
By default, the function returns the normalized fraction and the exponent as a two-element array. The normalized fraction and exponent satisfy the relation x = frac * 2^exp.
var pow = require( '@stdlib/math/base/special/pow' );
var x = 4.0;
var out = frexp( x );
// returns [ 0.5, 3 ]
var frac = out[ 0 ];
var exp = out[ 1 ];
var bool = ( x === frac * pow(2.0, exp) );
// returns true
If provided positive or negative zero, NaN, or positive or negative infinity, the function returns a two-element array containing the input value and an exponent equal to 0.
var out = frexp( 0.0 );
// returns [ 0.0, 0 ]
out = frexp( -0.0 );
// returns [ -0.0, 0 ]
out = frexp( NaN );
// returns [ NaN, 0 ]
out = frexp( Infinity );
// returns [ Infinity, 0 ]
out = frexp( -Infinity );
// returns [ -Infinity, 0 ]
For all other numeric input values, the absolute value of the normalized fraction resides on the interval [0.5,1).
frexp.assign( x, out, stride, offset )
Splits a double-precision floating-point number into a normalized fraction and an integer power of two and assigns results to a provided output array.
var Float64Array = require( '@stdlib/array/float64' );
var out = new Float64Array( 2 );
var y = frexp.assign( 4.0, out, 1, 0 );
// returns <Float64Array>[ 0.5, 3 ]
var bool = ( y === out );
// returns true
Notes
Care should be taken when reconstituting a double-precision floating-point number from a normalized fraction and an exponent. For example,
var pow = require( '@stdlib/math/base/special/pow' ); var x = 8.988939926493918e+307; // x ~ 2^1023 var out = frexp( x ); // returns [ 0.5000263811533315, 1024 ] // Naive reconstitution: var y = out[ 0 ] * pow( 2.0, out[ 1 ] ); // returns Infinity // Account for 2^1024 evaluating as infinity by recognizing 2^1024 = 2^1 * 2^1023: y = out[ 0 ] * pow( 2.0, out[1]-1023 ) * pow( 2.0, 1023 ); // returns 8.988939926493918e+307
Examples
var randu = require( '@stdlib/random/base/randu' );
var round = require( '@stdlib/math/base/special/round' );
var pow = require( '@stdlib/math/base/special/pow' );
var BIAS = require( '@stdlib/constants/float64/exponent-bias' );
var frexp = require( '@stdlib/math/base/special/frexp' );
var sign;
var frac;
var exp;
var x;
var f;
var v;
var i;
// Generate random numbers and break each into a normalized fraction and an integer power of two...
for ( i = 0; i < 100; i++ ) {
if ( randu() < 0.5 ) {
sign = -1.0;
} else {
sign = 1.0;
}
frac = randu() * 10.0;
exp = round( randu()*616.0 ) - 308;
x = sign * frac * pow( 10.0, exp );
f = frexp( x );
if ( f[ 1 ] > BIAS ) {
v = f[ 0 ] * pow( 2.0, f[1]-BIAS ) * pow( 2.0, BIAS );
} else {
v = f[ 0 ] * pow( 2.0, f[ 1 ] );
}
console.log( '%d = %d * 2^%d = %d', x, f[ 0 ], f[ 1 ], v );
}
C APIs
Usage
#include "stdlib/math/base/special/frexp.h"
stdlib_base_frexp( x, frac, exp )
Splits a double-precision floating-point number into a normalized fraction and an integer power of two.
#include <stdint.h>
double frac;
int32_t exp;
stdlib_base_frexp( 4.0, &frac, &exp );
The function accepts the following arguments:
- x:
[in] doubleinput value. - frac:
[out] double*destination for the normalized fraction. - exp:
[out] int32_t*destination for the integer power of two.
void stdlib_base_frexp( const double x, double *frac, int32_t *exp );
Examples
#include "stdlib/math/base/special/frexp.h"
#include <stdint.h>
#include <stdio.h>
#include <inttypes.h>
int main( void ) {
const double x[] = { 4.0, 0.0, -0.0, 1.0, -1.0, 3.14, -3.14, 1.0e308, -1.0e308, 1.0/0.0, -1.0/0.0, 0.0/0.0 };
double frac;
int32_t exp;
int i;
for ( i = 0; i < 12; i++ ) {
stdlib_base_frexp( x[i], &frac, &exp );
printf( "x: %f => frac: %f, exp: %" PRId32 "\n", x[i], frac, exp );
}
}