Electrical Engineering
      and Computer Sciences

Electrical Engineering and Computer Sciences

COLLEGE OF ENGINEERING

UC Berkeley

Combining Windows: A Performance Evaluation of Design Options

Philip Bitar

EECS Department
University of California, Berkeley
Technical Report No. UCB/CSD-93-727
February 1993

http://www.eecs.berkeley.edu/Pubs/TechRpts/1993/CSD-93-727.pdf

A combining window is an interval of time in a combining node during which incoming requests are gathered in the node in order to combine them into a single outgoing request. Our thesis is that a combining window is necessary in order to realize the dual forms of concurrency -- execution and storage concurrency -- that a combining tree is designed to achieve. Execution concurrency among the nodes at each level of a combining tree is necessary for the tree to achieve the speed up that it is designed to give. Without sufficient execution concurrency, the tree will not achieve the desired speed up. Storage concurrency among the nodes at each level of a combining tree is necessary for the tree to achieve the buffer storage that is required in order to implement the combining of requests. Without sufficient storage concurrency, node buffers will overflow. More specifically, the combining window shows how to bound node buffer size.

Advisor: Edward A. Lee


BibTeX citation:

@phdthesis{Bitar:CSD-93-727,
    Author = {Bitar, Philip},
    Title = {Combining Windows: A Performance Evaluation of Design Options},
    School = {EECS Department, University of California, Berkeley},
    Year = {1993},
    Month = {Feb},
    URL = {http://www.eecs.berkeley.edu/Pubs/TechRpts/1993/6017.html},
    Number = {UCB/CSD-93-727},
    Abstract = {A combining window is an interval of time in a combining node during which incoming requests are gathered in the node in order to combine them into a single outgoing request. Our thesis is that a combining window is necessary in order to realize the dual forms of concurrency -- execution and storage concurrency -- that a combining tree is designed to achieve. Execution concurrency among the nodes at each level of a combining tree is necessary for the tree to achieve the speed up that it is designed to give. Without sufficient execution concurrency, the tree will not achieve the desired speed up. Storage concurrency among the nodes at each level of a combining tree is necessary for the tree to achieve the buffer storage that is required in order to implement the combining of requests. Without sufficient storage concurrency, node buffers will overflow. More specifically, the combining window shows how to bound node buffer size.}
}

EndNote citation:

%0 Thesis
%A Bitar, Philip
%T Combining Windows: A Performance Evaluation of Design Options
%I EECS Department, University of California, Berkeley
%D 1993
%@ UCB/CSD-93-727
%U http://www.eecs.berkeley.edu/Pubs/TechRpts/1993/6017.html
%F Bitar:CSD-93-727