Computes exp(x) - 1 for small x.
A scalar, series or table.
A scalar, series, or table.
exp(1e-17) - 1
exp(log(1 + 1e-17)) - 1
For real values, EXPM1 is accurate for small x where 1 + x == 1 for double precision accuracy.
For small x, EXPM1(x) is approximately x, but EXP(x) - 1 is approximately 0 for double precision accuracy.
See LOG1P to compute LOG(1 + x) for small x, the inverse of EXPM1.