Karim Belabas on Mon, 09 Jan 2006 09:37:56 +0100


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

Re: length (lg) overflow


* buromano@rcn.com [2006-01-06 21:52]:
> I'm getting this message; length (lg) overflow,  using Dokchitser
> computel package  and can't seem to find a reference for it.  I image it
> relates to vector length.

PARI is trying to create a GEN object with length > 2^24 words [ e.g. a
vector with > 16777215 components ]

This is probably a sign of an inefficiency in computeL for your kind of
inputs, rather than a problem in PARI, but I would need a specific
testcase to be sure.

Hint: type 'trap' then trigger the problem and investigate within the
break loop the current values of the variables.

> Is  there any way to increase something to get around this?  Are there
> serious  penalties?

The easiest is to use a 64 bit machine. Another possiblity (if the
porblem indeed comes from large vectors) is to draw inspiration from

  http://www.math.u-bordeaux.fr/~belabas/pari/doc/faq.html#biggerlist

Cheers,

    Karim.
-- 
Karim Belabas                  Tel: (+33) (0)5 40 00 26 17
Universite Bordeaux 1          Fax: (+33) (0)5 40 00 69 50
351, cours de la Liberation    http://www.math.u-bordeaux.fr/~belabas/
F-33405 Talence (France)       http://pari.math.u-bordeaux.fr/  [PARI/GP]