Accumulator Architecture: An accumulator is anespecially designated register which supplies one instruction operand and receives result. Instructions in such machines are usually one-address instructions. The most common early architectures were DEC PDP-8, IBM 7090 etc.

On an Accumulator machine 'C = A + B'can be implemented as:

        LOAD A        // Load memory location A in accumulator

        ADD B          // Add memory location B to accumulator

        STORE C     // Store accumulator value in memory location C

Accumulator Architecture: Pros and Cons

• Implicit use of accumulator saves instruction bits
• Result is ready for immediate reuse however has to be saved in memory if subsequent computation doesn't use it right away.
• More memory accesses needed than stack. Consider a program to do expression:

        A = B * C + D * E.