Technical data

Vector Processing Concepts
1.1 SCALAR VS. VECTOR PROCESSING
Vector processing is a way to increase computer performance over that
of a general-purpose computer for certain scientific applications. These
include image processing, weather forecasting, and other applications
that involve repeated operations on groups, or arrays, of elements. A
vector processor is a computer optimized to execute the same instruction
repeatedly. For example, consider the process of adding 50 to a set of 100
numbers. The advantage of a vector processor is its ability to perform
this operation with a single instruction, thus saving significant processing
time.
In computer processors, a vector is a list of numbers, a set of data,
or an array. A scalar is any single data item, having one value. A
scalar processor is a traditional central processing unit (CPU) that
performs operations on scalar numbers in sequential steps. These types of
computers are known as single-instruction/single-data (SISD) computers
because a single instruction can process only one data item at a time.
A list of elements can be placed in an array. The array is defined by
giving each element and its location. Example:
12
is the value a located
at row 1 and column 2. The dimensions of the array are m and n. An
array element is a single value in an array, such as
12
below.
11 12 1
21 22 2
.
.
.
.
.
.
.
.
.
1 2
A one-dimensional array consists of all elements in a single row or single
column. A one-dimensional array can be expressed as a single capital
letter such as A, B, or C. Collectively, the elements within a vector are
noted by A(I), B(I), C(I), and so forth. Example: B and C are vectors,
where:
B = (-3, 0, 2)
The elements B(I) are –3, 0, and 2. The elements C(I) are 9 and –5.
1–2