Skip to main content

float64_add

binius_circuits::float64::add

Function float64_add 

Source 
pub fn float64_add(builder: &CircuitBuilder, a: Wire, b: Wire) -> Wire
Expand description

IEEE-754 double (binary64) addition built from small testable blocks. Returns a 64-bit IEEE-754 encoding.

Behavior summary:

  • Fully handles normals, subnormals, zeros (signed), infinities, NaNs
  • Rounding mode: round-to-nearest, ties-to-even
  • Sticky tracking across alignment and underflow right-shifts
  • Exact cancellation (+x) + (-x) → +0
  • Overflow → ±Inf with computed sign