float lower limit at given precision

[Jacopo Nespolo]

On 23 February 2012 23:38, Thomas D. Dean <tomdean at speakeasy.org> wrote:
> I assume by small, you mean close to zero? You want to do
>   if abs(n) < epsilon then n = 0"
>
> Look at
>   mpf_eq (mpf_t OP1, mpf_t OP2, mp_bitcnt_t op3)
> where mp_bitcnt is your needed precision.
>
> I used mpq_t for a large problem and at an intermediate point, set all n <
> value to zero.

Thanks, this looks like what I was looking for!
Thank you also for the other who answered.
Jacopo