From: Thomas Fuerle Newsgroups: comp.parallel.mpi Subject: Re: MPI and pthreads Date: Fri, 09 Apr 1999 13:00:20 +0200 Organization: Vienna University, Austria Message-Id: <370DDDC4.383805B7@vipios.pri.univie.ac.at> References: <7eiolf$4k9$1@finch.mathcs.emory.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Xref: ukc comp.parallel.mpi:4903 Paul Gray wrote: > I would appreciate any guidance on the feasibility of running > multi-threaded processes with MPI. I've been trying to write a > simple ping-ping program which utilizes one thread for sends and one > thread for receives, but have not been successful as of yet. > > I've tried numerous variations on pthread locks and signals to > prevent simultaneous calls to MPI routines -- combined with > nonblocking send and receives -- without much luck. AFAIK, mpich & pthreads are still using the same signal SIGUSR1. You must recompile mpich or pthreads, what you prefer with another signal, e.g SIGUSR2, SIGUNUSED (only on Linux), that they dont interfer with each other. The new mpich (1.1.1) version has a option LISTENERSIG, where you can define that signal, also the new LAM6.2beta. After recompilation mpich is still not MT-safe, so you must take care for critical sections on your own. hope this helps, Tom > > > Regards, > > Paul Gray > gray@mathcs.emory.edu > Math/Computer Science Department > Emory University, Atlanta, GA > > 150 North Decatur Building > 1784 N. Decatur Rd. > Atlanta, GA 30322 -- Thomas Fuerle mailto:fuerle@vipios.pri.univie.ac.at http://vipios.pri.univie.ac.at Institute for Applied Computer Science and Information Systems University of Vienna, Rathausstr. 19/4, A-1010 Vienna, Austria Tel: +43 1 4277 38423 --- Antoine de St. Exupery --- "Perfection is not achieved when there is nothing left to add, but when there is nothing left to take away".