On Friday, 8 August 2025 10:02:30 Pacific Daylight Time Tiago Freire wrote:
> No, _BitInt is a different philosophy of dealing with numbers. _BitInt is
> not multi-precision, it's fix precision but with support for widths larger
> than the CPU word size.
>
> And the compiler would not be able to optimize code the way multi-precision
> does because of limitations on what the compiler can reason about.

Why not? How would the division of two _BitInt(128) be different from what Hans
is proposing?

-- 
Thiago Macieira - thiago (AT) macieira.info - thiago (AT) kde.org
  Principal Engineer - Intel Platform & System Engineering