Conference rusure::math

1364.0. "Nice course for Vector support people" by CIVAGE::LYNN (Lynn Yarbrough @WNP DTN 427-5663) Fri Jan 04 1991 10:24

          Computational Math for Algorithmization & Vector Processing
                                 (EY-F607E-L0)


COURSE CODE:      EY-F607E-L0  SESSION NUMBER: 
LENGTH:           10 Days
TYPE:             Lecture/Lab

Dates:  February 18 - March 1, 1991	Location:  MRO2
	April 15-26, 1991			   MRO

Registration:  US-Field Area Registrars.  Europe & GIA-Send registration 
information to "HDQ REGISTRAR @OGO".

Intended Audience: Sales Support and EIS (PSS)

Course Description: 

This two week course will provide a foundation of computational mathematics as 
it relates to Vector and Parallel computing.  It incorporates aspects of 
numerical analysis, applied probability, applied mathematics, and computer 
architecture to provide an in-depth base for technical vector consulting. 

Instructor Background:

Ned Anderson

 B.Sc. MIT, Math. (Sloan National Scholarship)
 Ph.D. Royal Institute of Technology, Stockholm, Sweden.
 Thesis advisor Germund Dahlquist.
 Co-author and translator of internationally renowned textbook in
   Numerical Analysis (with Dahlquist, Bjorck).
 Author of 20 technical publications.
 Tenured Professor, Math./Comp.Sci., Kent State University, 1976-1984
 At DEC since 1984, working in computer arithmetic, approximation
   theory, standard functions, massive parallelism.
 Currrently Eng. Mgr., DXML (Digital Extended Math. Library)

Prerequisites: 

  *  Prerequisite: VAX Vector Processing for Programmers (EY-E197E-L0)
	aka VAX Vector Programming

  *  It is strongly recommmended that students review linear algebra and some 
	differential equations before attending.

  *  Background:

     -  Strong motivation to be involved in Vector machines or other areas
	which may involve compute intensive applications. 
     -  A science or technical background having completed a minimum of one 
	calculus course.
     -  Basic knowledge of linear algebra, differential equations, Fortran
        programming, and familiarity with issues in large scale scientific
        applications programs.


Course Objectives:  
  
Prepare Sales Support and EIS (PSS) personnel to provide technical vector 
consulting and the following vector processing support activities: 

  *  Determine if vector processing should be utilized to resolve the
     customer's problem 
  *  Develop in-depth familiarity with DXML vector/math library capabilities
  *  Translate algorithms into optimized vector programs
  *  Port and optimize code for vector processing
  *  Prepare course graduates to fulfill the functions of vector analyst and 
     programmers.
  
Course Outline:
  
	I.    General Principles of Numerical Computation
	II.   Obtaining and Estimating Accuracy
	III.  Approximation Of Functions
	IV.   Numerical Linear Algebra
	      A. Linear Systems
	          1. LU,QR factorizations
	          2. Sparse Methods
	          3. Iterative methods
	      B. BLAS, DXML, LAPACK
	      C. Eigenvalue Problems
	      D. Least Squares Problems, QR and Singular Value Decomposition
	V.    Root finding, Nonlinear Systems
	VI.   Differential Equations (ODE and PDE)
	      A. Finite Differences
	      B. Finite Elements
	VII.  Fast Fourier Transform and Applications
	VIII. Linear Programming and Optimization
	IX.   Parallel/Vector Computing, Algorithm/Programming Models
	X.    Application Focus (varies from course to course)
  
	      For the February 18th and April 15th courses the application 
	      focus will be Finite Elements.

Programming Exercises:  These will be heavily based on the use of DXML routines 
and LAPACK.  The application focus may involve the use of an additional program 
from that application area.

Hotel and Travel Information:

Will be sent through the registration system.  Flight reservations can be made 
through Logan Airport.

    Does a listing exist of all math courses that can be taken within DEC?
    
    Bo