Karim BELABAS on Wed, 12 Jun 2002 14:01:59 +0200 (MEST)


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

Re: 64bit HPUX binary buggy


On Tue, 11 Jun 2002, Igor Schein wrote:
> I compiled a 64bit gp binary on HPUX 11, but it's badly broken -
> a command as simple as bnfinit(x^2-2) keeps doubling the stack.
> I'd like to debug it, but I need some pointers on where to set
> breakpoints, what to look for, etc.

Well, you can hardly call 'bnfinit' a "simple" command.
Anything simpler breaking ? In 'make bench' for instance (avoiding the
'nfields' one), what benches are failing ? For what reason ?

Esp.

1) is basic arithmetic right, e.g
(13:54) gp > nextprime(10^9) \\ 1000000007

2) floating point ? (3.14, 1.2 * 3.4, sqrt(2), Pi, ...)

3) (tougher, but still sensible): polroots(), mathnf().

4) then nfinit(), and number field stuff. You can check that nfinit(x^2-2)
gives the exact same answer on your HP and on a better supported
architecture.

Try to evaluate what's (relatively) tough and what's easy, and send the
relevant _easy_ diffs from 'make bench' [e.g don't send a full bnfinit
output].

Then we can start investigating.

    Karim.

P.S: Btw, HP cc + portable kernel, I gather ?
-- 
Karim Belabas                    Tel: (+33) (0)1 69 15 57 48
Dép. de Mathematiques, Bat. 425  Fax: (+33) (0)1 69 15 60 19
Université Paris-Sud             Email: Karim.Belabas@math.u-psud.fr
F-91405 Orsay (France)           http://www.math.u-psud.fr/~belabas
--
PARI/GP Home Page: http://www.parigp-home.de/