Newsgroups: comp.parallel.mpi
From: Ahmad Afsahi <aafsahi@ece.uvic.ca>
Subject: MPI_Wtick and MPI_Wtime
Organization: University of Victoria
Date: Mon, 16 Mar 1998 12:26:44 -0800
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-ID: <350D8B04.B2B3A709@ece.uvic.ca>

Hi everybody,

I am trying to get some timing measurements of parallel
programs on network of workstations but I have understood that MPICH
dynamically gets the MPI_Wtick time (which perhaps affects the
MPI_Wtime)
so the timing might not be very precise, so I am considering to port my
applications
to IBM SP2 machines (to compare with). The question is that if I would
have the
same problem there on IBM SP2 with its own MPI-F or I would get precise
timing.

I appreciate any help.

Regards,
Ahmad Afsahi