Nothing Special   »   [go: up one dir, main page]
Scribd Logo
Upload

Welcome to Scribd!

  • 2 views

    ELC781 Assignment 1

    Uploaded by

    Omar Amer

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    ELC781 Assignment 1

    Uploaded by

    Omar Amer
    2 views1 page

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    2 views1 page

    ELC781 Assignment 1

    Uploaded by

    Omar Amer

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 1

    Cairo University Fall Semester Department of Electrical Engineering

    Faculty of Engineering 2024 – 2025


    Parallel Processing
    Professor: Mostafa Abd-El-Barr

    Assignments-1 (Due date: Noveber 3, 2022)


    1. Fill the following table
    Example Status Observation

    for(i=0; i<100; i++ )


    for(j=0; j<100; j++ )
    a[i][j] = f(a[i][j-1]);

    for( j=0; j<100; j++)


    for( i=0; i<100; i++)
    a[i][j] = f(a[i][j-1]);

    2. Conduct a comparative study between the following interconnection networks in terms of cost and performance:
    Interconnection Network Cost Performance
    The bus.
    The hypercube.
    The mesh networks.
    The fully connected.
    The multistage network.
    The crossbar network.

    3. Consider the two MINs shown below. Characterize these networks in terms of their ability to interconnect all inputs
    to all outputs simultaneously.
    0 0 0
    000 1 5 9 000
    001 1 1 1
    001
    2 2 2
    010 2 6 10 010
    011 011 3 3 3
    100 3 7 11 100 4 4 4
    101 5 5 5
    101

    110 110
    4 8 12
    6 6 6
    111 111 7 7 7

    4. Consider parallel architecture built using processors each capable of sustaining 0.5 megaflop operations. Consider a
    supercomputer capable of sustaining 100 megaflop operations. What is the condition (in terms of f) under which the
    parallel architecture can exceed the performance of the supercomputer?

    5. Consider a simple cost comparison between an n  n crossbar and an n  n Shuffle-exchange MIN. While the crossbar
    uses crosspoints, the Shuffle network uses 2  2 switching elements (SEs). Assume that the cost of a 2  2SE is four
    times that of a crosspoint. What is the relative cost of an n  n Shuffle-exchange network with respect to that of a
    crossbar of the same size? Determine the smallest value of n for which the cost of the crossbar is four times that of
    the Shuffle-exchange.

    6. Someone may argue against parallel processing saying that parallel processing is limited by the serial sections in the
    tasks. Is that true in your opinion? If NOT, can you argue against this by quoting scientific evidences and by quoting
    practical successful parallel processing experiences.

    Good Luck
    Page 1 of 1

    You might also like