bug-gmp
[Top][All Lists]
Advanced

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

Re: mpz_sizeinbase


From: Winfried Dreckmann
Subject: Re: mpz_sizeinbase
Date: Thu, 29 Mar 2001 08:00:51 +0200

Kevin Ryde <address@hidden> writes:

> Well, there should be 53 bits of precision on that log, which should
> be enough.
..
> I'm unable to reproduce that on a ppc604e (or an i386), I get ...78 as
> expected.  Is the compiler giving the right value for chars_per_bit?
>
>     printf ("%A\n", __mp_bases[10].chars_per_bit_exactly);
>
> should print 0X1.34413509F79FFP-2 by my reckoning.

I get 0X1.34413509F79FDP-2 on my machine which I think explains the
issue. For some reason 2 of those 53 bits get lost. But even with 53
bits of precision I see no proof that the original problem cannot occur.

> But your point is taken though, it'd be better not to depend on
> floating point for this.  A fixed point integer approximation would
> make it easier to guarantee the results, and would also hopefully make
> it possible to address the problem with 64-bit mpf exponents noted in
> doc/tasks.html.

The problem is that I want to use the function for a math library. So I
want it to be correct or, at least, know the range in which it is
correct.

Perhaps the following is helpful:

bash-2.03$ gcc -v
Reading specs from /usr/lib/gcc-lib/powerpc-suse-linux/2.95.2/specs
gcc version 2.95.2 19991024 (release)

bash-2.03$ uname -a
Linux tohu 2.2.14 #1 Tue May 30 01:15:50 GMT 2000 ppc unknown

bash-2.03$ ./config.guess
powerpc-unknown-linux-gnu

processor 604e
clock     170MHz
revision  2.2
machine   Power Macintosh

Winfried Dreckmann



reply via email to

[Prev in Thread] Current Thread [Next in Thread]