The accumulator architecture is an extreme compromise to the CISC ideal. It was used in the earliest computers, and again when computers were first implemented on a single integrated circuit.
In an accumulator architecture, there is only one register for data. it is called the accumulator. Arithmetic and logical instructions use the accumulator as both a source operand and as the destination operand. The third operand is in memory. These instructions require coding for the memory operand.
In addition, load and store instructions are needed to move data from memory into the accumulator or from the accumulator to memory. These instructions also require coding for the memory operand.
A = B*(C + D)
| Instruction | Code Size | Data | |
|---|---|---|---|
| load | C | 3 (1 + 2) | 4 |
| add | D | 3 (1 + 2) | 4 |
| mul | B | 3 (1 + 2) | 4 |
| store | A | 3 (1 + 2) | 4 |
| Total | 12 | 16 | |