Gerhard Niklasch on Wed, 4 Mar 1998 03:18:05 +0100

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

Re: dependence of format and realprecision

In response to
> Message-Id: <>
> Date: 	Wed, 4 Mar 1998 03:03:07 +0100
> From: Jack Fearnley <>
> I happen to be summing L-series to see if they sum to zero so I need a lot
> of internal precision but I do not need lots of decimal places cluttering
> up my screen.
> In GP-2alpha the format seems to be locked into the realprecision.  That is
> to say that when I change format the real precision changes accordingly.
> Am I missing something here or is there some reason why the previous
> flexibility is no longer provided?

But it is, provided that you chant the right incantations.

In 2.0 native mode (`compatible' set to zero):
(03:07) gp > default(realprecision ,100)
   realprecision = 100 significant digits
(03:07) gp > default(format)  
   format = g0.100
(03:07) gp > default(format,"g0.20")
   format = g0.20
(03:07) gp > Pi
%1 = 3.1415926535897932384
(03:07) gp > \x
[&=001a7010] 0500000d  40800001  c90fdaa2  2168c234  c4c6628b  80dc1cd1
  29024e08  8a67cc74  020bbea6  3b139b22  514a0879  8e3404dd  ef9519b3  
(so increasing the precision does change the format, but there is no
problem changing the format afterwards to whatever may be convenient).

Passing to full compatibility mode:
(03:09) gp > default(compatible,3)
  ***   Warning: user functions re-initialized.
   compatible = 3 (use old functions, ignore case )
(03:12) gp > \precision=100
   realprecision = 100 significant digits
(03:12) gp > \format=g0.20
   format = g0.20
(03:12) gp > Euler
%2 = 0.57721566490153286060
(03:12) gp > \x
[&=001a7e10] 0500000d  407fffff  93c467e3  7db0c7a4  d1be3f81  0152cb56
  a1cecc3a  f65cc019  0c03df34  709affbd  8e4b59fa  03a9f0ee  d0649ccb  

Suggestion for the future:  This kind of question and response would
have given some life to the pari-users list...

Shameless Plug(tm):  Since you're posting this from Concordia, Montreal
and I'm typing from the U Laval campus -- fancy coming over to Quebec to
our little Pari-2.0.6.alpha demonstration tomorrow morning at 09:30 am,
Pavillon Alexandre Vachon, room #1610?  (In broken Franglais, I haven't
quite mastered the Quebecois accent yet! ;^)  Otherwise, see you in
the number theory seminar on Thursday...

Enjoy, Gerhard