Course Name:

Course Title: Special Topics in High-Performance Grid Computing and Research Networking

Prerequisite:

Admission to the Global CyberBridges fellowship program (see www.cyberbridges.net for application information).
Familiarity with sequential programming in at least one of the major programming languages such as C, C++, Java, FORTRAN, and the like.

Course Goals:

This course provides students with knowledge of the fundamental concepts of high performance computing as well as hands-on experience of the core technology in the field. The lectures will cover topics such as High-Speed Networks, Design and Implementation of Clusters, Globus Toolkits, Virtual Data Systems, Resource Scheduling, MPI, OpenMP, Web Services, Grid Services, and Job Flow Management. This course will include lectures, seminars, and a course project. The course project will require the students to "gridify" a scientific application. Students will conduct experiments on the CyberBridges cluster, and leverage extramural resources, such as LA Grid, TeraGrid, and Open Science Grid as requirements might justify.

Topics:

HPC Concepts:

Concurrency

Concurrent, Parallel, Distributed?
Reasons for Concurrency
History of Concurrency

Concurrent Computers

Within a CPU, within a “box”, or across Boxes?
Taxonomy

Modeling and Design

A Parallel Machine Model
A Parallel Programming Model
Partitioning, Communication, Agglomeration, and Mapping

Performance

Time, Rate, “Peak”, Benchmarks, GHz, Speedup
Amdahl’s Law
Efficiency, Scalability, Bottleneck, Latency
Profiling

HPC Hardware, Software, and Networking:

Networking

Ethernet physical components: Copper, Fiber, Switches
IP networking: Address space, Routing
Issues in HPC: Bandwidth Delay Product, Frame size, Latency/Jitter
Design of cluster networks

Cluster Computing

Hardware: Architecture, Planning, Installation
Cluster Software: Rocks, MPICH, Local schedulers

Grid Computing

Standards: OGSA, OGSI, WSRF
Toolkit: GT4, GSI, GRAM, MPICH-G2, Condor

HPC Programming:

Cluster Programming

MPI
Open MP

Grid Programming

Web Services
Grid Services

References:

Designing and Building Parallel Programs: Concepts and Tools for Parallel Software Engineering by Ian Foster.

Available online at URL: http://www-unix.mcs.anl.gov/dbpp/

Computer Networks, 4th Edition, by Andrew S. Tanenbaum, Prentice Hall (ISBN: 0130661023).
High Performance Linux Clusters with OSCAR, Rocks, OpenMosix, and MPI By Joseph D. Sloan, First Edition November 2004.
Globus® Toolkit 4: Programming Java Services (The Morgan Kaufmann Series in Networking), by Borja Sotomayor, Lisa Childers.
Kevin Dowd et.al, "High Performance Computing", O'Reilly Press
Wiliam Gropp et.al, "Using MPI Portable Parallel Programming with the Message-Passing Interface", MIT Press
Thomas Sterling, "Beowulf Cluster Computing with Linux", MIT Press

Other reading material: Class notes.

Grading Policy:

Class Attendance and Participation: 20%.
Project: 40%.
Technical Paper: 40%.

Grading Standard: The grading scale is: A: 90 | A-:87 | B+:84 | B: 80 | B-:77 | C+:74 | C: 70 | C-:65 | D+:60 | D: 55 | D-:50. Note that a B- is not a B.

Attendance: Attendance may be taken during each class meeting.

Some important dates:

February 6 Tuesday Last day to withdraw from the University with a 25% refund of tuition.
March 12 Monday Last day to drop a course with a DR grade. Last day to withdraw from the University with a WI grade.
March 19 - 24 Spring Break.
April 19 Thursday Classes end.
April 20 - 21 Fri-Sat Exam Study Days (No exams given on these days).
April 23 - 28 Mon - Sat Official Examination Period.
April 30 - May 1 Mon- Tues. Commencement Exercises.
May 3 Thursday Complete grade report available to students by web and at kiosks.
For more information, please consult the university calendar at: http://www.fiu.edu/~register/calendars.htm

Policy on Make-up Examinations and Assignments: There will be no make-up exams. All project deliverables and assignments should be submitted at the beginning of class on the due date. The only excuse for missing an exam is verifiable cases of illness and emergencies and religious holy days. Please check the dates of the exams and inform me of any conflicts with religious holy days as soon as possible.

Code of Academic Integrity:

http://www.fiu.edu/~oabp/misconductweb/2codeofacainteg.htm

University Policies: academic misconduct, sexual harassment, religious holydays, and information on services for students with disabilities.

http://www.fiu.edu/provost/polman/sec2/sec2web2-44.htm

http://www.fiu.edu/~eop/EOPSexH.pdf

http://www.fiu.edu/~provost/polman/sec19web.html

Last modified 3/28/2007