Re: Google Summer of Code proposal - libm audit

[Joerg Sonnenberger <joerg%britannica.bec.de@localhost>]

On Tue, Mar 30, 2010 at 08:25:02PM +0300, Stathis Kamperis wrote:
> 1. A certain number of unit tests for every math function. The ATF
> would fit nicely here. As part of last year's gSoC, I've written some
> math test cases, e.g. [1]. Such tests reveal bugs that would otherwise
> go unnoticed, e.g. [2], [3]. We will target correctness, precision,
> edge case stressing, error handling and every other requirement of the
> IEEE 754 standard we will be confronted with. We can even build a
> database of known-good-results from well established CAS's and match
> it against the output of our libm.

It would be nice to have exhaustive testing for float values against a
known-to-be-good implementation (e.g. an implementation of perfect
rounding) as well as randomised testing against such code. This should
enforce the maximum documented ulp, but allow any value in that delta.

> 2. Improvements on existing bits (such as error handling) or
> implementation for missing functions.
> 3. Some sort of documentation of the algorithms nature, efficiency, behavior.

Agreed. For long double support you are aware of the different
precisions? I think at least two formats are present ATM, but I am not
sure about a third version (80bit and 128bit for sure). Do you intend to
deal with the specifics of VAX (and may be Alpha)?

> Regarding the politics.

Ignore them :)

Joerg