Information Technology Reference
In-Depth Information
M
M
M
M
M
M
P
P
P
P
P
P
( a )
( b )
p
p
p
p
p
p
p
p
MMMMMMMM
( c )
( d )
Figure 11.1 Four types of multiprocessor interconnection networks. (a) Single-bus system,
(b) multi-bus system, (c) multi-stage interconnection network, (d) crossbar system
a classification scheme for computer architecture should provide a basis for infor-
mation ordering and a basis for predicting the features of a given architecture.
Two broad schemes exist for computer architecture classification. The first is
based on external (morphological) features of architectures and the second is based
on the evolutionary features of architectures. The first scheme emphasizes the
finished form of architectures, while the second scheme emphasizes the way an
architecture has been derived from its predecessor and suggests speculative views
on its successor. Morphological classification provides a basis for predictive
power, while evolutionary classification provides a basis for better understanding
of architectures. Examining the extent to which a classification scheme is satisfying
its stated objective(s) could assess the pros and cons of that scheme.
A number of classification schemes have been proposed over the last three dec-
ades. These include the Flynn's classification (1966), the Kuck (1978), the Hwang
and Briggs (1984), the Erlangen (1981), the Giloi (1983), the Skillicorn (1988), and
the Bell (1992). A number of these are briefly discussed below.
11.2.1. Flynn's Classification
Flynn's classification scheme is based on identifying two orthogonal streams in a
computer. These are the instruction and the data streams. The instruction stream
is defined as the sequence of instructions performed by the computer. The data
stream is defined as the data traffic exchanged between the memory and the
processing unit. According to Flynn's classification, either of the instruction or
data streams can be single or multiple. This leads to four distinct categories of
Search WWH ::




Custom Search