Re: precision of result when adding t_REAL with lots of cancellation

Jeroen Demeyer on Wed, 12 Feb 2014 13:16:02 +0100

Re: precision of result when adding t_REAL with lots of cancellation

To: pari-dev@pari.math.u-bordeaux.fr
Subject: Re: precision of result when adding t_REAL with lots of cancellation
From: Jeroen Demeyer <jdemeyer@cage.ugent.be>
Date: Wed, 12 Feb 2014 13:15:49 +0100
Cc: denis.simon@math.unicaen.fr
Delivery-date: Wed, 12 Feb 2014 13:16:02 +0100
In-reply-to: <r9wr478wmw2.fsf@host-56-227.warwick.ac.uk>
References: <r9wr478wmw2.fsf@host-56-227.warwick.ac.uk>
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Thunderbird/24.2.0

On 2014-02-12 13:12, Peter Bruin wrote:

Would it be reasonable to round the precision _down_ in such cases?

If you would that, then you would very quickly end up with a precisionof 0 bits after a few such additions.

Follow-Ups:
- Re: precision of result when adding t_REAL with lots of cancellation
  - From: Peter Bruin <P.Bruin@warwick.ac.uk>

References:
- precision of result when adding t_REAL with lots of cancellation
  - From: Peter Bruin <P.Bruin@warwick.ac.uk>