Modern Math Expm1

Home › Math Library › Math Library › Expm1

Expm1

Returns e raised to a power, minus one, accurately for powers near zero.

Syntax

Expm1( x )

Notes

• For a tiny x, exp( x ) - 1 subtracts two nearly equal numbers and loses most of its digits. Expm1 keeps them.
• Log1p is its inverse.
• Overloaded on float64 and on float32: a single-precision argument takes the single-precision path and does not widen.
• BOUNDED accuracy: correct to the platform library's stated ulp bound.

Examples

#import "std/numeric"
using numeric

print Expm1( 1e-10 )         '' 1.00000000005e-010
print exp( 1e-10 ) - 1.0     '' 1.000000082740371e-010 -- wrong in the 8th decimal place
print Expm1( 1.0 )           '' 1.718281828459045

---

Library: Math Library
Category: Math Library
Documented under: The libm surface