Newsgroups: comp.parallel.pvm
From: sdblackb@uncc.edu (Stuart D Blackburn)
Subject: Re: PvmDataInPlace takes longer than PvmDataRaw???
Organization: University of NC at Charlotte
Date: 9 Aug 1995 15:32:07 GMT
Message-ID: <40akdn$99s@news.uncc.edu>

In article <404f70$bua@ds2.acs.ucalgary.ca>, adilger@enel.ucalgary.ca (Andreas Dilger) writes:
> In article <3vtqmn$qu7@news.uncc.edu>,
> Stuart D Blackburn <sdblackb@uncc.edu> wrote:
> [snip]
> >PVM Version 3.2 output  
> >	* 20x10x10 matrix with 0.01 tolerance
> >	* Using PvmDataInPlace encoding
> >
> >Slave 1 finished 76 iterations in 5.3855 seconds
> >==============================================================
> >PVM Version 3.1 output  
> >	* 20x10x10 matrix with 0.01 tolerance
> >	* Using PvmDataRaw encoding
> >
> >Slave 1 finished 76 iterations in 2.5458 seconds
> [snip]
> 
> I've been looking at this, an I wonder if the output you're giving is
> actually from 2 differenct versions of PVM, or is this coincidentally the
> version numbers of your software?  I'm guessing PVM versions here, so I
> wonder if you have tried running with PVM 3.2 and PvmDataRaw?  It is
> possible that there is something else with the PVM code that is slowing
> things down.  
	Sorry about that. I should have been more clear about that. The version
number shown is the version of my software. 
	I am going to look into Philip Papadopoulos' suggestion that PvmDataInPlace
is causing a separate packet to be sent for each pvm_pkXXX statement. This would
definately explain the overhead that I seem to be seeing.
	If this is indeed the way in which PvmDataInPlace is working (atleast on workstation clusters), then this should be better documented. 

	Thanks for all of the suggestions.

	Stuart
> 
> Cheers, Andreas.
> -- 
> Andreas Dilger    University of Calgary   \ "If a man ate a pound of pasta and
> (403) 220-8792    Micronet Research Group  \  a pound of antipasto, would they
> Dept of Electrical & Computer Engineering   \    cancel out, leaving him still
> http://www-mddsp.enel.ucalgary.ca/People/adilger/          hungry?" -- Dogbert





