Getting Started on the Intel/Paragon

<title>Getting Started on the Intel/Paragon</title>
<h2>Getting Started on the Intel/Paragon</h2>
<hr><pre>

		 INTRODUCTION -- I/Paragon processing

I/Paragon Access & Setup:
-------------------------
Once a successful logon to cs.sandia.gov has been accomplished, you will
need to modify your .cshrc file in the following manner:
 
set IPARAON_USER = yes            # Do you use the I/Paragon system

This variable is originally set to no.  If it is not set to yes, your 
environment will not allow I/Paragon usage.

From cs.sandia.gov, you have access to the I/Paragon system by directly
logging into the machine itself; it is called "acoma".

-------------------------------------------------------------------------
|	Using the machine - quick overview
-------------------------------------------------------------------------

It is requested that acoma be used to only load and debug programs.
Compilation of programs should be done on one of the "compute servers";
either jemez or pajarito.  Setting the IPARAGON_USER variable
to "yes" will provide access to all of the necessary software.  Some of
the more frequently used commands are:

     if77	Compile a Fortran program
     icc	Compile a C program
     iCC	Compile a C++ program
     pexec	Used for node acquisition
                  Example: % pexec -sz 16  /* this acquires 16 ndoes */
     pman	man pages for the I/Paragon

-------------------------------------------------------------------------
|	Who to contact for help						|
-------------------------------------------------------------------------

If you have questions about the I/Paragon machine, please Email them to:
		iparagon@cs.sandia.gov
This is a distribution list so that several knowledgable people may answer
the question.

Sandia has two on-site Intel/SSD representatives.  You can contact during
normal working hours (8:00am to 5:00pm MT):
  Mike Proicou	mcproic@cs.sandia.gov	505-845-7998
    Applications level support
  Mack Stallcup	tmstall@cs.sandia.gov	505-845-7998
    Hardware and OS level support

There is also a mailing list of the current I/Paragon users at:
		iparagon-users@cs.sandia.gov

-------------------------------------------------------------------------
|	Using the machine - remote compilation				|
-------------------------------------------------------------------------

It is suggested that the operating environment for the I/Paragon machine
be treated similarly to the I/Gamma machine.  Specifically, you use a
Sun machine for cross-compiling your programs; either jemez or pajarito.
You then log into the I/Paragon machine to do the actual load/running of
the programs.  Internal sandians can have their workstation setup so that
they can cross-compile on their own machine.

You will have to modify your existing ~/.cshrc file to incorporate
the new paths used to support Sun cross-compilation.  The subset .cshrc
template is found in /Net/local/sys/Cshrc.common.  Look at the
IPARAGON_USER section for the relevant paths.  It is included below:
[ New users are already setup for this. ]

if ( ! ${?IPARAGON_USER} ) then
    set IPARAGON_USER = no
endif
if ( "$IPARAGON_USER" == "yes" ) then
    setenv PARAGON_XDEV /usr/iparagon/current
    setenv NX_DFLT_PART .compute
    setenv TTY `tty` >& /dev/null
    set path = ($path ${PARAGON_XDEV}/paragon/bin.sun4)
    alias pman "man -M ${PARAGON_XDEV}/paragon/man"
endif

Also note, that if you are both a IGAMMA_USER and a IPARAGON_USER, there
are major conflicts in the cross-development paths.  EG, you use 'icc' for
C cross-compilation for both the I/Gamma and I/Paragon machines.  I have
yet to come up with a good solution to this problem; suggestions are
welcome.  The only "solution" I have so far is that users should have two
shell scripts, eg dogamma and doparagon, that each contain the relevant
sections of the subset .cshrc.  If you work on the I/Paragon machine you
run %doparagon first which alters the $path variable so that the I/Paragon
icc comes first in the search list.  Similarly for I/Gamma work.

-------------------------------------------------------------------------
|	Using the machine - native load/run				|
-------------------------------------------------------------------------

The name of the I/Paragon machine is 'acoma'.  When you login to acoma,
you will be deposited onto the .service partition.  From there you can
load/run programs.  The network file systems are NFS mounted onto acoma,
so there should be no problems with cross-compiling on a Sun.

You will have to modify your existing ~/.cshrc file to incorporate
the new paths used to support Sun cross-compilation.  The master .cshrc
template is found in /Net/local/sys/Cshrc.  Look at the "OSF1"
section for the relevant paths.  It is included below:

...
else if ( "$uname" == "Paragon OSF/1" ) then
# Welcome to a I/Paragon front-end

set path=(/usr/bin /usr/sbin /sbin . /usr/local/bin)
setenv NX_DFLT_PART .compute

else
...
<hr>
</pre>