CIS 570 Advanced Computer Systems.

Homework Assignments  Lecture Notes    Projects     Exams   

Instructor: Dr. Boleslaw Mikolajczak

Prerequisite: CIS 273,CIS 370 or equivalent knowledge

Time: Tu, Th 6.30 - 7.45 pm, Dion 101

Office hours: Fr 1-3pm, Dion302F AND Tu, Th 9.00 am -10.00 am in Dion 302F
bmikolajczak@umassd.edu
phone: (508)-999-8350
fax: (508)-999-9144

Textbook: “Computer Architecture, A Quantitative approach” by D.Petterson, J. Hennessy,Morgan Kaufmann publishers , Third edition, 2003

COURSE OBJECTIVES:

Centered around the concepts of quantitative analysis of computer systems and parallel/distributed processing
with four architectural classes: pipeline computers, vector processors, array processors, and multiprocessor systems

COURSE DESCRIPTION:

This course provides students with optimization of the classical von Neumann architecture and different forms of
parllel architecture.The course coverage includes classification of parallel computer systems, memory and input
/output subsystems, principles of pipelining and vector processing, pipeline computers, vectorization methods,
SIMD and MIMD computers, and multiprocessors architectures
Course Description PDF

REFERENCE BOOKS:

1.High Performance Computer Architecture, H.S.Stone,Addison-Wesley,1993,third edition
2.Advanced Computer Architecture: Parallelism, Scalability ,Programmability,by K Hwang,McGraw-Hill,1993.
3.UNIX Systems For Modern Architectures.Symetric Multiprocessing and caching for Kernal Programmers,
C.Shimmel,Addison-Wesley,Professional Computing Series,1994.

MAJOR TEXTBOOK CHAPTERS COVERED IN THIS COURSE:

1.Fundamentals of Computer Design
2.Instruction Set Principles
3.Pipelining
4.Advanced Pipelining and Instruction-Level Parallelism
5.Memory-Hierarchy Design
6.Storage Systems
7.Interconnection Networks.
8.Multiprocessors.

SOME DETAILED PROBLEMS AND ISSUES DISCUSSED:

  • CRAY T3D- shared memory MPP- course contents by example of modern handout parallel architecture and problems to be solved .Trio:application ,architecture (hardware,software),algorithm.
  • Classification of parallel computer architectures;Parallel computer models
  • Program and network properties for parallelism
  • Principles of scaleable performance,Speedup performance laws:Amdahl's and Gustafson's laws.Isoefficiency analysis
  • Memory subsystems and their design,cache memory,principles of locality
  • Buses :Cache memory technology;interleaved memories,virtual memory technology;coherency/consistency models
  • Pipelined processors: linear,non-linear pipeline design,arithmetic pipeline,superscalar and design superpipelinedesign
  • Cache coherency protocols and synchronization mechanisms
  • Message passing mechanisms,routing algorithms, deadlock and virtual channels
  • Vector processing and vector computers.SIMD architectures.The Connection machine CM-5
  • Latency hiding techniques.MultiThreading.Scaleable and multi threaded architectures
  • Parallel programming models,languages,programming,Dependency analysis,code optimization,scheduling,loop
    parallelization,Parallel program development and environments

    COURSE GRADING POLICIES :

    Division Points  Number Value
    MIDTERM 20 1 20
    Final exam 20  1 20
    Homework 8  4 32
    Programming Assignments 14  2 28
        Total  100

     

    EVALUATION CRITERIA:

  • A+(Points 98-100)
  • A(Points 94-97)
  • A-(Points 91-93)
  • B+(Points 87-90)
  • B(Points 84-86)
  • B-(Points 81-83)
  • C+(Points 78-80)
  • C(Points 74-77)

    • COURSE PRINCIPLES:

    Project and homework are due on the date specified and are expected to be done individually by students