CS 415/515: High Performance Computing
Spring 2008
Syllabus

Course description: This course covers the ways that High Performance Computing (HPC) techniques can be used to address problems in Computational Science. After introducing the major scientific applications areas and basic concepts of parallel computing, it outlines the hardware design of modern HPC platforms and the parallel programming models that they support. The principal methods of measuring and characterizing serial and parallel performance are then covered. The final section of the course gives an introduction to grid technologies together with an overview of the opportunities and challenges associated with the computational grid.

Pre-requisites: CS 310 (Algorithm Analysis), CS 350 (Networking) and MATH 247 (Linear Algebra)
 

Course Website
http://bates.cs.mnsu.edu/cs415
Check the website regularly for announcements and updates. 

Course Hours and Location

Lectures: MW 3:30-4:45pm

WH 286 or WH 119

Office Hours

Monday	Tuesday	Wednesday	Thursday	Friday
by appointment	1-2:30, 4:30-5:30	1:30-2:30*	1-2:30*	2-4**

* in WH 283
** in ACC or WH 231

If things that are useful for the entire class come up, they will be posted on the announcement section of the class webpage so please check it regularly.

Course Materials
Required Text: Parallel and Distributed Programming Using C++, Cameron Hughes & Tracey Hughes, Addison Wesley, 2004.
Additional readings will be distributed in class or via the website.
Recommended Texts: An Introduction to Parallel Computing: Design and Analysis of Algorithms, Ananth Grama, George Karypis, Vipin Kumar, Anshul Gupta, Second Edition, 2003.
A list of related text books will be made available.

Student Outcomes

Students who complete this course will be able to:

1. Describe operational aspects of several parallel computing architectures.
2. Read and understand programs written in a language specific to High-Performance computing.
3. Write programs that use concurrency in the form of multiple threads.
4. Write programs that use the Message Passing Interface (MPI).
5. Measure, analyze and assess the performance of HPC programs.
6. Evaluate the suitability of different HPC solutions to standard problems in Computational Science.
7. Describe the potential benefits and issues of grid computing.

Added Expectations of Graduate Students

1. Graduate students will be held to a higher standard in all coursework, including assignments and exams.
2. Graduate students will also be expected to perform in depth and thorough independent investigation of the subject matter.

Tentative Topic List

1. Introduction to HPC/linux & the field of Concurrent Computing (~ 1 wk)
2. Shared-Memory Computing: (~ 4 wks)
   a. Multiple Processes / Synchronization
   b. Locks and Barriers
   c. Semaphores
   d. Monitors
3. Distributed-Memory Computing: (~ 3 wks)
   a. Message Passing
   b. Remote Procedure Call
   c. Distributed Computing Paradigms
4. Parallel/Distributed Applications (~ 3 wks)
   a. Grid Computations
   b. Matrix Computations
5. High-Performance Computing Tools (~ 3 wks)
   a. Compilers
   b. Libraries
   c. Development Tools

Grading Distribution
Homework and Lab work: 15%
2 Midterm Exams: 40%        
Final project: 45%

Important Dates
Midterm 1: February 27
Midterm 2: April 16
Final Project Presentations: Last week of course
Final Project Presentations: May 9

Course Tools

• HPC & Krypton: You will get an account on both of these machines.  If you haven't set up a krypton account, please do so since that will be the portal to the HPC (High Performance Computer).  You can set up the account through the mavmail/mavdisk account and password pages.
• Desire2Learn: This web-based software package will be used for turning in programming assignments and feedback surveys.
• Password protected webpages: The Computer Science Lab computers host these pages.  The assignments and solutions will be available on the web but will be password protected.
• Supplemental materials and information will be made available on the course website.