TorchLean API

FloatLib.Floats.Formats.Posit.Arithmetic.Limb.Packed.Mul.Runtime

Direct packed-pair posit multiplication runtime #

The packed-pair multiplication kernel returns posit codes for configured values stored in two native limbs. Range and semantic refinement proofs live in Mul.Proof.

@[noinline]

Decode two packed operands, multiply their significands in four native limbs, and return the complete result encoding in two native limbs.

The exact product is normalized with a sticky low bit before rounding. Significands, the product, and the result code stay in fixed-width carriers; exponents and bit counts use Int and Nat.

Instances For