Newsgroups: comp.parallel.pvm,comp.parallel.mpi
From: eddemain@neumann.uwaterloo.ca (Erik Demaine)
Subject: Re: Update: Conversion MPI to PVM
Organization: University of Waterloo
Date: Mon, 4 Mar 1996 22:31:32 GMT
Message-ID: <DnrL8K.66F@undergrad.math.uwaterloo.ca>

a psychedelic psychopath (jay%cs.wmich.edu) wrote:
: ...
: one thing that we had to do however, was change all Bsend's to 
: Send's.  Thus, the implementation possibly existant non-block aspect
: is lost.

This seems odd.  What implementation of MPI are you using?

I can't say I recommend Bsends, since the amount of buffer space is hard to
determine before program execution.  If you can use Isend, you should get
even higher performance.

: a "quick" question, how does one force a buffered non-blocking send/recv?
: or should i read before i ask?

MPI_Ssend

Ciao!
Erik
--
Erik Demaine                 \/  e-mail: eddemain@neumann.uwaterloo.ca
Dept. of Computer Science    /\  URL   : http://barrow.uwaterloo.ca/~eddemain/
University of Waterloo       \/  Most jugglers do poorly while drunk, especially
Waterloo, ON Canada N2L 3G1  /\  in a large wavelength gravity field.

