[PATCH 05/15] benchtests: Add cospif

[DJ Delorie]

Adhemerval Zanella Netto <adhemerval.zanella@linaro.org> writes:
>> I assume the comment is wrong, since cospi(x)'s domain is unbounded, and
>> a few multiples of pi seems reasonable.
>
> You are right, I used the same input range CORE-MATH uses for performance
> evaluation and it is on the range of [-4,4].  I will update the comment.

With that update, LGTM :-)

Reviewed-by: DJ Delorie <dj@redhat.com>