Module precision 

Floating-point comparison and rounding utilities.

This module provides tools for safe floating-point arithmetic, including multiple comparison strategies (absolute, relative, ULP-based) and rounding utilities.

Constants

EPS

Machine epsilon for f64 (2^-52).

F32_EPS

Machine epsilon for f32 (2^-23).

Functions

abs_diff

Absolute difference: |a - b|.

almost_eq

Absolute-tolerance equality: |a - b| <= eps.

almost_eq_rel

Relative-tolerance equality: |a - b| <= tol * max(|a|, |b|). Equal when both are zero.

almost_eq_ulp

ULP-based floating-point equality for f64.

ceil_to_multiple

Round x up to the nearest multiple of m.

copysign

Copy the sign of b to the magnitude of a.

epsilon

Machine epsilon for a generic Real type.

floor_to_multiple

Round x down to the nearest multiple of m.

is_close

Combined absolute + relative tolerance comparison.

is_infinite

Check if f64 is infinite (positive or negative).

is_nan

Check if f64 is NaN.

is_subnormal

Check if f64 is subnormal (denormal).

next_float

Next representable f64 toward positive infinity.

prev_float

Next representable f64 toward negative infinity.

relative_diff

Relative difference: |a - b| / max(|a|, |b|).

round_to

Round x to decimals decimal places.

round_to_multiple

Round x to the nearest multiple of m.

significant_figures

Round x to n significant figures.

ulp

Distance from x to the next representable f64 toward +infinity.