From: Alexandre Carissimi Newsgroups: comp.parallel.mpi Subject: Re: Start Threads in MPI Date: Tue, 15 Sep 1998 20:40:25 +0200 Organization: LMC-Laboratoire de Modelisation et Calcul Message-Id: <35FEB499.149B9B65@imag.fr> References: <35EFE6EF.F62842F7@csis.hku.hk> <35FE4C66.B8080E64@vipios.pri.univie.ac.at> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Thomas Fuerle wrote: > > Zheng Yu Liang wrote: > > > Hi, > > > > We need start some p_threads in mpi process. But when we start mpi with > > two nodes , only process 0 enter the first thread and halt when the > > thread return . Anyone know why? And can I call a mpi method in a > > thread? > > > > Our workstation is SMP with Linux 2.0.x , and mpich 1.1.0. Following is > > our program: > > > > Hi, i'm fighting with the some similar problems in Linux. > I use now SIGUSR2 instead of SIGUSR1 in mpich 1.1.1 and if the below output > is the right output, we made it. Hi; You have TWO problems: 1. When you use linuxthreads do you have a signal conflit as said before. You need to change SIGUSR1 and SIGUSR2 to some other non used signals. You can find all steps on linux sites to do that. 2. MPICH, if compiled to use P4 as communication library, doesn't work correctly. P4 is not thread-safe. I recommend you to use LAM .Until now I didn't have problem with LAM and threads using redhat5.1 (the more importante is to use glibc-2.0) and LAm 6.1 (with patch level 7). Regards, ASC __________________________________________________________________ CARISSIMI, Alexandre Alexandre.Carissimi@imag.fr Laboratoire de Modelisation et Calcul - LMC Institut National Polytechnique de Grenoble - INPG 100, Rue des Mathematiques Fone: (33) 04.76.51.46.03 38041 - Grenoble - Cedex 9 FRANCE Fax : (33) 04.76.51.48.46 (International calls removes 0 from 04) "Le seul moteur pour gagner c'est innover" Renault (F1 - 1996) ________________________________________________________________