forum@abinit.org
Subject: The ABINIT Users Mailing List ( CLOSED )
List archive
- From: Yann Pouillon <pouillon@pcpm.ucl.ac.be>
- To: forum@abinit.org
- Subject: Re: [abinit-forum] parallel version
- Date: Thu, 24 Mar 2005 12:27:46 +0100
- Organization: Physico-Chimie et Physique des Matériaux
Dear Manuela,
I answered to your question a little bit too quickly last week. Here are
some more precisions about the parallel version of ABINIT.
Actually, parallelism in ABINIT is not that feudal (master-slave). It
follows a SIMD model, i.e. all processes perform the same task and one
of them eventually gathers the data before displaying or saving it.
I'm not sure exactly what you mean by "peer-to-peer". If this implies
fault-tolerance, i.e. the execution continuing smoothly even if one node
crashes, please note that ABINIT is not-at-all fault-tolerant in this
respect.
Hope this make things a little bit clearer.
Best regards,
Yann.
--
Yann Pouillon UCL/SE/FSA/MAPR/PCPM
Tel: +32 (0)10 47 81 91 Bâtiment Boltzmann, a+1
Fax: +32 (0)10 47 34 52 1 place Croix du Sud
Mel: pouillon@pcpm.ucl.ac.be 1348 Louvain-la-Neuve (Belgique)
- parallel version, manuela . volpe, 03/17/2005
- Re: [abinit-forum] parallel version, Yann Pouillon, 03/18/2005
- Re: [abinit-forum] parallel version, Yann Pouillon, 03/24/2005
Archive powered by MHonArc 2.6.16.