[libre-riscv-dev] correctly-rounded fp functions algorithm

[Luke Kenneth Casson Leighton]

On Tue, Oct 22, 2019 at 9:26 PM Jacob Lifshay <programmerjake at gmail.com> wrote:
>
> I had an idea:
> when we're trying to make a implementation of a correctly-rounded fp
> function, we should be aiming to approximate the rounded results rather
> than the exact mathematical result, since that allows us to have better
> approximations with more wiggle room in the approximations, since the
> result we want is the rounded result, not the exact mathematical result.

yes.  like the way that sin is constructed (as a taylor series), you
can stop at some point because further values are never going to give
you better than what the mantissa can store.

e.g. well.. just like how CORDIC operates.

> thought I should write this all down before I forget.

:)