Harvard architecture

Harvard architecture is a computer design with separate storage and communication paths for instructions and data. It allows instruction fetching and data access to occur simultaneously.

Instruction memory: Memory that stores the instructions a processor executes. It supplies instructions over a path distinct from data access.
Von Neumann architecture: A computer architecture in which instructions and data share memory and communication paths. It is the standard contrast to Harvard's separate instruction and data paths.
Microcontroller: An integrated circuit containing a processor, memory, and peripherals for embedded control. Many microcontrollers use Harvard-style paths to fetch code while handling data.
Harvard Mark I: An electromechanical computer completed at Harvard University in 1944, also known as the IBM Automatic Sequence Controlled Calculator. Its separate instruction and data media helped inspire the architecture's name.
Data memory: Memory that stores values a processor reads and writes while executing instructions. Its separate path can carry data while instruction memory supplies the next instruction.
Princeton architecture: A computer organization that stores instructions and data in a shared memory. The name is often used for the unified-memory alternative to Harvard organization.
Digital signal processor: A processor designed to perform repeated numerical operations on digital signals efficiently. Separate instruction and data access can sustain continuous signal-processing workloads.
Howard Aiken: An American physicist and computer pioneer who led the development of the Harvard Mark I. Aiken's machine embodied an early separation of instructions and data.
Instruction fetch: The processor operation that retrieves the next instruction for execution. A dedicated instruction path can fetch while data transfers proceed.
Unified memory architecture: An arrangement in which processors use a common memory space for different kinds of data or access. A unified memory space differs from Harvard's distinct instruction and data storage.

Linked from 7 pages