COM 1400 Data Parallel Computing

Course Format
Course Coordinator
Textbooks and References

Prerequisites by Topic
Major Topics Covered in Course
Laboratory Projects

Course Description

Introduces the basic concepts of parallel computer architectures, network topologies, and data parallel programming.  Emphasizes SIMD machines (with mesh or hypercube interconnections) and networks of workstations.  Studies fundamental data structures and data parallel algorithms for matrix operations and fast fourier transforms.  Examines graph and geometric computations for complexity and performance characteristics.  Requires a significant amount of programming to complete course assignments.

4 QH credit
Prerequisite: COM 1390, MTH 1125, MTH 1301.

Course Information

Course is offered only during the Spring quarter. CS majors are guaranteed a place in class.
 

Spring 2xxx (This course will not be offered in the academic year 2000-2001)

Course Format

This is an elective course for  BS CS major and a  'focused elective' course for BA CS majors.

Course Coordinators

Professor Gene Cooperman
gene@ccs.neu.edu

Textbooks and References

Spring 2000

• Parallel Programming:  Techniques and Applications Using Networked Workstations and Parallel Computers, by Barry Wilkinson and Michael Allen.

References:
 
• Course materials handed out

Course Goals

Cheap, powerful parallel computers exist today.  One can program in threads on top of shared memory computers, as witnessed by the multi-processor offerings of UNIX and Windows-based vendors.  One can also program distributed, socket-based applications for workstations over Ethernet using both RISC and Pentium architectures.  Higher parallel abstractions are being built on top of this framework.  The course provided an overview of many of these concepts from both a software/hardware and theoretical viewpoint.  Software topics included:  POSIX threads, sockets, MPI, CILK (for SMP), QUARK (for DSM), TOP-C (for NoW). and Legion (for wide area distributed computing).  Theoretical topics included parallel prefix, pointer jumping, parallel sorting and parallel merging, Brent's Theorem, PRAM and circuit models and bridging models (QSM, BSP, logP).

Prerequisites by Topic

COM1390, MTH1125, and MTH1301 (algorithms, calculus 3, and linear algebra 1);  Beyond that I required computer science maturity with a good command of C.

Major Topics Covered in the Course

Software topics included:  POSIX threads, sockets, MPI, CILK (for SMP), QUARK (for DSM), TOP-C (for NoW). and Legion (for wide area distributed computing).

Theoretical topics included: parallel prefix, pointer jumping, parallel sorting and parallel merging, Brent's Theorem, PRAM and circuit models and bridging models (QSM, BSP, logP).

Laboratory projects

Course project of student's choosing.