Indirection in IAS computer

Question

From Computer Organisation and Architecture:

The IAS operates by repetitively performing an instruction cycle. Each instruction cycle consists of two sub cycles. During a fetch cycle, the opcode of the next instruction is loaded into the IR and the address portion is loaded into MAR. This instruction may be taken from the IBR, or it can be obtained from memory by loading a word int he MBR, and then down to the IBR, IR, and MAR.

Now here, is there part I do not understand:

Why the indirection? These operations are controlled by electronic circuitry and result in the use of data paths. To simplify the electronics, there is only register that is used to specify the address in memory for a read or write and only one register used for the source or destination.

Can anyone explain this to me more intuitively? Why is this indirection being used?

Answer 1

the book is describing low level machine language architecture associated with the CPU & what is also known as "microcode". it would be possible to design opcodes that use multiple registers (that does occur in some mostly uncommon cases) but most opcodes generally refer to one register or a memory location, unless its a transfer between two registers. the register number is usually encoded into the opcode in binary. that is, registers serve as a sort of small intermediate "hub" or even a small "cache" between memory and the CPU.

if one had opcodes that refer to multiple registers it quickly leads to a combinatorial explosion where the opcodes become very numerous. even a 2-way register opcode scheme takes roughly $c \cdot r^2$ encodings where $r$ is the number of registers. a 3-way register operation would require $c \cdot r^3$ etc... another scheme is to limit multiway register operations to just a few "privileged" registers, not the full set.

a good case study or example might be the 6502 CPU that has an accumulator, an auxiliary X and Y register, and some limited operations to move between them. a full/general level of register interchange is then supported by stringing several/multiple register manipulation opcodes together.