Russ Cox • 1/19/2026

Fast Unrounded Scaling: Proof by Ivy

This article presents a detailed mathematical proof and algorithm for 'fast unrounded scaling,' a core operation in floating-point to decimal conversion. It defines the unrounded form, explains the 'Scale' algorithm using 128-bit approximations, and proves its correctness for use in fixed-width, short, and parse conversion algorithms as part of a larger series on simple and fast floating-point formatting.

0 comments

#floating-point #algorithm #Scaling