[Beowulf] Bcast or reduce?

Anthony Skjellum tony at mpi-softtech.com
Thu Aug 26 14:38:19 PDT 2004

Bcast would let you do 1-to-all sharing of the array from the root
MPI_Allreduce or reduce could be finagled into concatenating disjoint sets
of primes BUT
   MPI_Allgatherv seems better for concatenating disjoint sets of numbers,
preceeded by an MPI_Allgather to share the lengths of the disparate sizes.

What processes know what about what primes?  Where are new primes figured
out?  IF in all processes, then incremental Allgatherv's might help.


----- Original Message ----- 
From: "Jack C" <jack at crepinc.com>
To: <beowulf at beowulf.org>
Sent: Thursday, August 26, 2004 3:42 PM
Subject: [Beowulf] Bcast or reduce?

> Hey,
> My project (as some of you may remember from my last post) is a prime
> basically just so that I can learn MPI. I've gotten to the point where
> split up and runs great: the only thing left is that I'm testing each
> number's divisibility agaist ALL the numbers between 2 and sqrt(number).
> Obiously, this isn't needed.
> What I need to do is have an array of integers full of my previously found
> primes to test against. However, I'm having problems figuring out how to
> this running tab updated on all the nodes.
> Will MPI_Bcast or _reduce help? (just thought I'd through that out
> Thanks guys,
> -Jack C
> jack {at} crepinc.com
> http://www.crepinc.com
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org
> To change your subscription (digest mode or unsubscribe) visit

More information about the Beowulf mailing list