Subject: Re: [boost] [Review Request] Multiprecision Arithmetic Library
From: Christopher Kormanyos (e_float_at_[hidden])
Date: 2012-04-03 15:52:18
> > > Just a crazy idea ... what about if the exponent is a big integer as
> > > well? And the mantissa ... guess you can go all crazy, hope this is
> > > supported by the library!
> > My main point against this idea would be that this might drop performance a
> > lot.
> I think, that with speicalization that should at least be possible have
> special, high performance implementations for e.g. int32 and int64 as
> exponent. With the type of the exponent as a template parameter that sould
> work well.
Well, I like it... And I don't like it---right now.
The increased testing effort would throw this library back, maybe,
almost a year (at least my float back-end part of it). I tested that
back-end with automatically generated test cases from Mathematica
in combination with state-of-the-art code coverage off and on for the
better part of a year. Realistically, one would recommend keeping
a flexible exponent on the *very interesting* list, yet moving on with
the tested implementation that we have today. If the exponent were
to be templated in the future, a default template parameter could
be used to retain back compatibility.
Best regards, Chris.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk