From: lusk@donner.mcs.anl.gov (Rusty Lusk)
Newsgroups: comp.parallel.mpi
Subject: Re: MPICH on SNI RM-600 (?)
Date: 4 Nov 1998 15:47:17 GMT
Organization: Argonne National Laboratory
Message-Id: <71psu5$e1k@milo.mcs.anl.gov>
References: <71kpuo$q0n$1@ns.mtu.ru>


In article <71kpuo$q0n$1@ns.mtu.ru>, "Alexander N. Andreyev" <alexander@vvv.srcc.msu.su> writes:
|> Hello, MPICH developers and users.
|> Have somebody ever tried to build MPICH on an RM-600 system from Siemens
|> Nixdorf, running their Reliant Unix? If yes, please share your experience. I
|> see that RM-600 is not on the list of MPICH's directly supported systems;
|> but, as an SMP system, it must be supported through ch_shmem device. My
|> questions are: 1) which configuration should I use for building MPICH on an
|> arbitrary Unix working on SMP system? 2) if I need to modify MPICH
|> distribution for this platform, what are general directions of such
|> modification?

MPICH uses a configure script to try to identify features of its environment
at run time.  In theory, you should be able to do

  configure -device=ch_shmem >& configure.log
  make      >& make.log

and the configure script will try to make intelligent
guesses about what makefiles should look like to build MPICH on your system.
Whether this works in practice depends on just how wierd the system is.
The log files will identify the problems.  If they are incomprehensible, send
them to mpi-bugs@mcs.anl.gov

Regards and good luck,
Rusty Lusk