CS240A: Applied Parallel Computing: Winter 2006

Interdisciplinary introduction to applied parallel computing on modern supercomputers. Topics include applications oriented architecture, software systems, MPI, data parallel systems, parallel MATLAB, caches and vector processors, and parallel algorithms for numerical algorithms such as linear algebra, N-body problems, multigrid, fast multipole, wavelets and Fourier transforms, mesh generation, and partitioning. The course includes a project with hands-on emphasis on understanding the realities and myths of what is possible on the world's fastest machines.

Class Meeting

Instructor

Assistant

Discussion and Announcements

The mailing address for just the instructor and assistant is cs240a@cs.ucsb.edu.

Course Requirements

• Programming homeworks (probably 3): 45%
• Final project: 45%
• Class participation: 10%

Homework

• Homework 1 (due Mon Jan 30): Problem, FAQ, Sequential Matlab code, UPC solution
• Homework 2: (due Fri Feb 10): Problem, FAQ, Sequential Matlab code
• Homework 3: (due Wed Mar 8): Problem, FAQ, Sequential Matlab code

Final Project

• Links to project web pages

• Due dates and requirements

• Final project demo schedule (Mon Mar 20)

Documentation and Machine Setup Information

Visit the documentation and machine information page. (This is now the current, 2006 version. It will get more updates as the class goes on.)

Class Schedule and Lecture Slides

• Mon Jan 9: Introduction [lecture slides] [Mark Adams talk]

• Wed Jan 11: Parallel programming language intro: MPI and UPC (Viral) [Code][MPI Guide][UPC talk]

• Mon Jan 16: (Martin Luther King Day, no class.)
  (Homework 1 assigned)

• Wed Jan 18: Models of parallel programming and architecture [lecture slides] [HPCS study]

• Mon Jan 23: Matrix multiplication, cache, and memory locality [lecture slides]

• Wed Jan 25: Communication and computation cost modeling (Rob Schreiber, HP Labs)

• Mon Jan 30: Matlab*P (Viral)
  (Homework 1 due, Homework 2 assigned)

• Wed Feb 1: MathWorks' Parallel Matlab (Cleve Moler, The MathWorks)

• Mon Feb 6: Parallel dense matrix algorithms [lecture slides]

• Wed Feb 8: Parallel sparse matrix algorithms [my slides, Kathy Yelick's slides]
  (Homework 2 due FRI FEB 10)

• Mon Feb 13: Trees, scans, PRAMs, and log n [lecture slides, HPCS defect study]
  (Homework 3 assigned)

• Wed Feb 15: The Cray MTA: A novel shared-memory architecture (John Feo, Cray Inc.)
  (Final project topics and teams due)

• Mon Feb 20: (Presidents' Day, no class.)

• Wed Feb 22: Parallelism in physical simulation I [lecture slides, sharks and fish]

• Mon Feb 27: Parallelism in physical simulation II [same slides as Feb 20, Robert Funnell's intro to finite elements]

• Wed Mar 1: Solving systems of linear equations [my slides, Jim Demmel's slides on dense Gaussian elimination, Sherry Li's slides on SuperLU]
  (Final project progress report due)

• Mon Mar 6: Sparse linear systems, graph partitioning [my slides on Fiduccia-Mattheyses, Alan Edelman's slides on spectral and geometric methods, Kathy Yelick's overview slides]

• Wed Mar 8: Parallel graph algorithms [SSCA2 slides, web page; papers on maximal independent sets and connected components]
  (Homework 3 due)

• Mon Mar 13: Discussion

• Wed Mar 15: Final project presentations

• Mon Mar 20: Final project demos in GSL

• Wed Mar 22: (Final project paper due)

Interesting Links

• Top 500 SuperComputers
• Top 10 Algorithms
• Dead SuperComputer Society
• Twelve Ways to Fool the Masses When Giving Performance Results on Parallel Computers
• A nice introduction to finite element methods in a biomedical application
• Horst Simon's Feb 2005 CSE panel talk [PPT] [PDF]

Similar Courses Elsewhere and Elsewhen

• Applied Parallel Computing, MIT (link)
• Applications of Parallel Computers, UC Berkeley (link)
• CS240A, Spring 2005, UCSB (link)