Project

General

Profile

Bug #267

Cyclotomic(106743) fails

Added by John Abbott over 11 years ago. Updated almost 11 years ago.

Status:
Closed
Priority:
High
Assignee:
John Abbott
Category:
Incomplete function
Target version:
CoCoA-5.0.3
Start date:
18 Oct 2012
Due date:
% Done:

100%

Estimated time:
Spent time:
1.00 h

Description

How is this possible?

# Cyclotomic(106743,x);
ERROR: Inexact division (i.e. quotient does not exist in ring or monoid)

I only wanted to verify some results in my BigFactor paper then CoCoA-5 does this to me? :-(
[moral of the story: don't verify your results prior to publishing :-]

Related issues

Related to CoCoALib - Design #268: Exponent range (in power products)Closed2012-10-18

Related to CoCoALib - Slug #679: power for PPs is too slowClosed2015-04-13

History

#1 Updated by John Abbott over 11 years ago

  • Status changed from New to Resolved
  • Assignee set to John Abbott

The problem was silent exponent overflow. In the file config.H there is the line:

  typedef unsigned short SmallExponent_t;

Anna found the clue when gcd(x^106743-1,x^4641-1) gave x-1 which is impossible because 106743=23*4641. It was an easy step to guess that exponent overflow had occurred.

So what should the definition in config.H be? The one which uses less space, is fine for almost all computation, and is probably a bit faster (i.e. unsigned short); or one which allows "ridiculously large" computations (i.e. unsigned int, or should we even go all the way to unsigned long or unsigned long long for users of C++11?). How ridiculous should we allow the user to be? :-)

#2 Updated by John Abbott over 11 years ago

  • Status changed from Resolved to Closed
  • % Done changed from 0 to 100

Closing this issue because the problem is that addressed in issue #268 (namely the range of exponents to allow).

#3 Updated by Anna Maria Bigatti almost 11 years ago

  • Category set to Incomplete function
  • Target version set to CoCoA-5.0.3

Also available in: Atom PDF