MathGroup Archive 2004

[Date Index] [Thread Index] [Author Index]

Search the Archive

FactorIntegerECM repeated or parallel runs question

  • To: mathgroup at smc.vnet.net
  • Subject: [mg46871] FactorIntegerECM repeated or parallel runs question
  • From: dont at agora.rdrop.com (Don Taylor)
  • Date: Fri, 12 Mar 2004 23:39:13 -0500 (EST)
  • Sender: owner-wri-mathgroup at wolfram.com

If I understand the documentation in V5 for FactorIntegerECM
correctly, it always seeds the random number generator with
a fixed value and this means that running this on several
different machines or running this several times on the same
machine when hunting for an elusive factor will just duplicate
the identical calculations, rather than start from different
values with one lucky run perhaps finding a factor.

Is my understanding of this correct?  If so, is there a way to
subvert this and get separate runs to do independent work?

I'm looking for factors of some large numbers, with little
chance of success.  But if it were easy to start a fresh
run of this every time I knew the machine was going to be
idle for a day or two then a miracle might happen some night.

If my understanding of this is correct, is there some reason
for this that could be revealed?  Or perhaps that this might
be changed with an option in a future version?

Thanks for all your support


  • Prev by Date: Re: Undiscovered Bug?
  • Next by Date: Re: Re: Creating a symmetric matrix
  • Previous by thread: Re: AES implementation in Mathematica
  • Next by thread: Reduce