add "workload" traces for sin/exp/pow in binary32/64/128

[Adhemerval Zanella]

On 26/08/2020 10:46, Adhemerval Zanella wrote:
> 
> 
> On 26/08/2020 05:06, Paul Zimmermann wrote:
>>        Dear Adhemerval,
>>
>>> LGTM, thanks.  There is just a nit on benchtests/sinf-inputs with a gratuitous new line.
>>
>> sorry for that, here is a new version without that gratuitous new line.
>>
>> Please can someone push it for me, since I have no commit access?
> 
> I will push it for you.

Sorry for the delay, I push it upstream.  I have also split the patch, one for
each symbol (as we discussed privately).