Ilya Zakharevich on Mon, 18 Dec 2023 09:06:40 +0100


[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

Re: libpari precision handling changes 

On Mon, Dec 18, 2023 at 12:14:23AM +0100, Bill Allombert wrote:
> > > Now the precision is a multiple of BITS_IN_LONG.
> > 
> > Is not there the code 'b' for this?  Why not use it?!

> It is not the same. 'b' allows any number of bits, while 'p' allows only multiple of
> BITS_IN_LONG. This is necessary until we allows t_REAL with precision in bits.

Hurray Hurray HURRAY!!!

Thanks,
Ilya

  P.S.  Is there a theory how to implement this to (statistically)
  	minimize the precision drift in long chains of calculations?