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

Re: MPI vs PVM

>>>>> "junichi" == Junichi Uekawa <dancer@netfort.gr.jp> writes:

    junichi> My impression is that PVM is a old project that is dying a death.

Alternatively, that it is a mature project that just "works", and like
an old grandmother, is reliable and un-sexy.  There is nothing like a
product that "almost works right" to get people involved; once
involved, the project evolves...

    junichi> We have two competing implementations of MPI (mpich and lam)
    junichi> and they are both still alive, and scalability issues are
    junichi> being worked on.

Re: MPICH, sometime I'm going to have to file a bug so that it doesn't
require rsh style services, but will let one use ssh...  (that is the
main reason I've been avioding MPICH backend development).

    junichi> At supercomputing conferences, my impression is that PVM is no
    junichi> longer actively developed, and MPI is trying to replace it.

and has been trying for years...

Speaking of replacements, has anyone looked into packaging any of the
public compute grid infrastructure toolkits for debian, i.e. Globus or
made an attempt at HARNESS, or similar (I can't recall if Legion
distributed source; I'm sure there are others I'm not aware of; this
isn't my primary research area).

A.J. Rossini				Rsrch. Asst. Prof. of Biostatistics
U. of Washington Biostatistics		rossini@u.washington.edu	
FHCRC/SCHARP/HIV Vaccine Trials Net	rossini@scharp.org
-------------- http://software.biostat.washington.edu/ ----------------
FHCRC: M: 206-667-7025 (fax=4812)|Voicemail is pretty sketchy/use Email
UW:   Th: 206-543-1044 (fax=3286)|Change last 4 digits of phone to FAX
(my tuesday/wednesday/friday locations are completely unpredictable.)

Reply to: