Harvardarkkitehtuuria

Harvardarkkitehtuuria is a computer architecture concept that separates instruction and data pathways. This means that instructions and data can be fetched simultaneously, leading to faster processing speeds compared to a Von Neumann architecture, which uses a single pathway for both. In a Harvard architecture, there are two distinct memory spaces, one for instructions and one for data, each with its own bus. This separation allows for parallel access.

The concept originated at Harvard University, hence the name. Early microcontrollers and Digital Signal Processors (DSPs) commonly employed this architecture due to its efficiency in handling real-time signal processing tasks where the speed of instruction and data fetching is critical. Modern processors often use a modified Harvard architecture, where the instruction and data caches are separate, but they share a common main memory. This hybrid approach aims to retain the benefits of parallel access at the cache level while maintaining the simplicity of a unified main memory. The primary advantage of the Harvard architecture is its potential for increased throughput and reduced instruction cycle time. However, it can also lead to increased hardware complexity and memory management challenges due to the separate memory spaces.