LAM features extensive debugging support in the application development cycle and peak performance for production applications. LAM features a full implementation of the MPI communication standard.
LAM is portable to all UNIX machines and includes standard support for SUN (SunOS and Solaris), SGI IRIX, IBM AIX, DEC OSF/1, HPUX, and LINUX.
subscribe lam
in the body of the message (the contents
of the subject line are irrelevant).
This mailing list is archived at
http://www.mpi.nd.edu/MailArchives/lam/.
Instructions on unsubscribing from the list can be found in the
FAQ.
A complete implementation of the MPI-1.1 specification is part of LAM.
- Quick Tutorials
- MPI Quick Tutorials give the new user a fast start.
- MPI Quick Reference Card, PostScript, two sided fold-up, 8.5 x 11
- Standard Document
- The MPI standard and other documents are available for browsing or download at the MPI Forum.
- Miscellaneous
- An MPI FAQ is maintained at Mississippi State University.
- Notre Dame maintains a comprehensive list of MPI implementations.
- Review the results of the MPI Poll '95.
- Top 10 reasons to prefer MPI over PVM.
- Bibliography and Papers
- ORNL maintains a bibliography of MPI related papers.
- "Robust Message Delivery Through Guaranteed Resources" describes implementation and portability issues in core MPI message delivery. It is of special interest to implementors.
- "The Performance of LAM 6.0 and MPICH 1.0.12 on a Workstation Cluster"
- Other MPI sites on WWW
%cat lamhosts # a 4-node LAM # By removing the # sign you can boot up to 6 machines. #ibm7 140.117.35.37 #ibm8 140.117.35.38 #ibm9 140.117.35.39 #ibm10 #140.117.35.40 #ibm26 # You may not want to include ibm26, like I said it is a "lemon". #140.117.35.86 #ibm27 140.117.35.87Each machine will be given a node identifier (nodeid) starting with 0 for the first listed machine, 1 for the second, etc.
The recon tool verifies that the cluster is bootable.
% recon -v lamhosts recon: testing n0 (140.117.35.37) recon: testing n1 (140.117.35.38) recon: testing n2 (140.117.35.39) recon: testing n3 (140.117.35.87)The lamboot tool actually starts LAM on the specified cluster.
% lamboot -v lamhosts LAM 6.0 - Ohio Supercomputer Center hboot n0 (140.117.35.37)... hboot n1 (140.117.35.38)... hboot n2 (140.117.35.39)... hboot n3 (140.117.35.87)...lamboot returns to the UNIX shell prompt. LAM does not force a canned environment or a "LAM shell". The tping command builds user confidence that the cluster and LAM are running.
% tping -c1 N 1 byte from 4 nodes: 0.154 secs 1 message, 1 byte (0.001K), 0.154 secs (0.013K/sec) roundtrip min/avg/max: 0.154/0.154/0.154
% hcc -o hello hello.c -lmpi
% mpirun -w n0-3 /Your/full/directory/hello From process: 0 out of 4, Hello World! From process: 1 out of 4, Hello World! From process: 2 out of 4, Hello World! From process: 3 out of 4, Hello World!You have seen a print out from 4 different nodes.
Congraduations!!
% lamclean -v killing processes, done sweeping messages, done closing files, done sweeping traces, done
% wipe -v lamhosts tkill n0 (140.117.35.37)... tkill n1 (140.117.35.38)... tkill n2 (140.117.35.39)... tkill n3 (140.117.35.87)...
Now you have killed the LAM session on the network.
Acknowledgements
This page is mostly the same as the page created by LAM / MPI Parallel Computing
Laboratory for Scientific Computing
University of Notre Dame
The LAM Team
A specical thanks to Jeff Squyres, he helped me a lot during the installation, also S.Y. Wang spent hours helping me to read through the installation guide.
Chieh-Sen Huang
Office: 2002-2
e-mail: huangcs@math.nsysu.edu.tw
URL: http://www.math.nsysu.edu.tw/u/huangcs