to the contents of register 2, R2. By changing the value
in R1, different operands may be addressed. This is
particularly useful for addressing memory locations in
arrays. Indexing is a very useful troubleshooting tool.
A short routine can be written to form a program loop
to isolate on a specific malfunction.
On some computers, a CPU register is dedicated to
this indexing function. In those cases, it is called an
index register and is usually 3-bits or more depending
on the computer type.
Some computers permit a
general-purpose register to be used as an index register.
Indirect Operand Addressing
Indirect addressing enables the operand address to
vary during program execution by specifying a location
in memory or a register in the instruction that tells where
the address will be stored. See figure 8-15. In this
example, the braces are used to tell that register 2 has
been specified to hold the operand address. This means
the contents of the main memory location whose
address is contained in R2 are added to the contents of
R1. Like the indexed mode of addressing, the indirect
mode of addressing offers flexibility and is useful in
addressing an array of data. Because the actual address
pointing to an array can be stored separately from the
program in memory, a large number of array pointers
can be used.
Relative Operand Addressing
In many computers, particularly those with
multiprogramming capabilities, a separate set of
registers called base registers is used to define the start
of particular blocks or segments of memory. Each
block of memory could contain a separate application
program. The contents of a base register is called the
base address.
Any execution of instructions or
referencing of operands within the block of memory
defined by the base requires that an offset or relative
address be used. The offset is added to the base during
instruction execution to allow reference of the correct
instruction or operand address.
INSTRUCTION SIZE
Each address of memory (main or ROM) contains
a fixed number of binary positions or bits. The number
Figure 8-15.—Example of indirect addressing.
of bits stored at a single address varies among types and
generations of computers. For example, some store 8
bits (1 byte) at each location; others store 16, 32, or more
bits at each location. The size of each memory location
or memory word has a direct effect on the execution of
machine instructions.
Basic instructions deal with full word exchanges
as the register size is usually the same as the memory
word size. In most computers, particularly those with
large memory words, the capability exists to transfer
less than a full memory word of information between
memory and the applicable register.
This allows
memory words and registers to be further divided into
economically sired bit groups for the most efficient use
of memory for information storage and handling. For
example, it is preferable to store two 8-bit characters in
one 16-bit memory location than to waste an extra
16-bit location for the second character. Let’s examine
some of the various instruction sizes.
Full- or Single-Word Instructions
A full- or single-word instruction simply uses all
the data contained in the instruction word to execute the
instruction regardless of the size: 8-bit, 16-bit, and so
on. Refer back to figures 8-3 and 8-5 for examples of
full- or single-word instructions, 16-bit and 32-bit.
Half-Word Instructions
Half-word (upper or lower half) instructions consist
of one-half of the normal instruction word size. The
half-word instructions are executed by acquiring the
complete normal instruction word, consisting of the
half-word instruction to be executed and the next
sequential instruction.
After the first half-word
instruction is executed, it is followed by the execution
of the next sequential half-word instruction. If only one
half-word instruction is used, it is usually located in the
upper half of the instruction word with all zeros in the
lower half of the instruction word. Refer back to figures
8-6 and 8-7 for examples of a half-word instruction.
Character-Addressable Instructions
In computers with word lengths greater than 8 bits,
character-addressable instructions allow specified bit
fields (called characters) of a word to be processed by
the instruction. This is done in lieu of processing a
whole-, half-, or quarter-word operand. Character
addressing is permitted only when the instruction is
executed in the indirect address mode. The particular
operand bit field to be acquired is specified by the
8-12
