
		      Portable MPI Model Implementation

        		  Version 1.2.0, December 1999

                                 David Ashton
                                 Anthony Chan
				  Bill Gropp
				  Rusty Lusk
                                 Debbie Swider
                                 Rajeev Thakur
			 Argonne National Laboratory


Major Changes

This release provides full MPI 1.2 compliance, including cancel of sends.
It also contains many bug fixes, particularly for LINUX systems.
Other major changes include:

o IMPI (Interoperable MPI) style flow control.
o A Windows NT version is now available as open source.  
o Support for SMP-clusters in mpirun.
o A Fortran 90 MPI module (actually two modules)
o Support for the routine MPI_INIT_THREAD (but only for MPI_THREAD_SINGLE) 
o Support for VPATH-style installations, along with a installation
  process and choice of directory names that is closer to the
  GNU-recommended approach
o A new, scalable log file format, SLOG, for use with the MPE
  logging tools.  SLOG files can be read by a new version of Jumpshot
  which is included with this release.
o A new version of ROMIO
o A new device for networked clusters, similar to the p4 device but based
  on daemons and thus supporting a number of new convenience features,
  including fast startup. 

There have been a number of changes to the configure, build, and installation
process, particularly when building for multiple devices and/or
architectures.  Please read the installation manual in mpich/doc/install.ps.gz
for details.  

Known problems and bugs with this release are documented in the file
mpich/KnownBugs .  

There is an FAQ at http://www.mcs.anl.gov/mpi/mpich/faq.html .  See this if
you get "permission denied" when trying to run MPICH.

A change log for all changes from the previous release is available from
http://www.mcs.anl.gov/mpi/mpich/r1_2_0changes.html .

There is a paper on jumpshot available at
ftp://ftp.mcs.anl.gov/pub/mpi/jumpshot.ps.gz .

What's next:

We continue to enhance MPICH with a goal of full MPI-2 capability.  In
addition, we continue to explore ways to improve the performance of MPICH.  In
the queue are improved datatype handling and a new sockets implementation.  In
the longer term, a new ADI with better support for multithreaded applications
and new network approaches will be introduced.  One area where we'd like
contributions is better implementations of the collective routines.  Please
contact us if you'd like to contribute (send mail to gropp@mcs.anl.gov).


Reporting Bugs
==============
If you have trouble, first check the installation and user manuals (in 
mpich/doc/{guide,user}.ps.Z) .  Next, check the on-line bug list at
http://www.mcs.anl.gov/mpi/mpich .  Finally, if you are still having problems,
send

   The type of system (often, uname -a will give you this)
   The output of configure
   The output of make
   Any simple programs or tests

to mpi-bugs@mcs.anl.gov .  Do NOT use the newsgroup comp.parallel.mpi ; this
newsgroup is for discussions about MPI, not any particular implementation of
MPI.

