Multi-core and Parallel Systems

Introduction

• Central Processing Unit (CPU): The brain of a computer responsible for executing instructions.
• Early computers had a single CPU on a physical chip for all instructions.
• Evolution led to multi-core processors and parallel systems due to increased demands on the CPU.

Multi-core Processors

• Definition: Single physical chip containing two or more independent processing units (cores).
  • Each core can fetch, decode, and execute its own instructions.
• Chip Multiprocessor (CMP): A more accurate term for what is commonly called a CPU today.
  • CMP contains multiple CPU cores.
  • Examples:
    • Dual-core: 2 cores
    • Quad-core: 4 cores
    • New processors can have up to 64 cores.
• Performance Considerations:
  • Enhanced by on-chip shared cache and communication methods.
  • Doubling cores doesn't necessarily double performance due to overheads and software limitations.
  • Multi-core systems generally process multiple instructions faster than single-core systems.

Parallel Processing

• Definition: Dividing program instructions among multiple processors or cores to increase speed.
• Objective: Run programs in less time by:
  • Assigning separate processes/cores to different stages of the fetch-decode-execute cycle.
  • Splitting a single job into tasks executed on different units.
• Considerations:
  • Speed improvement depends on the task and software design for parallel processing.

Amdahl's Law

• Purpose: Predicts theoretical speed increase in task execution by adding resources.
  • Used in parallel computing to predict speed increases from more processors/cores.
  • Simply adding more cores may not always lead to significant speedup.

Parallel vs. Concurrent Processing

• Concurrent Processing: System can support multiple actions in progress at the same time.
• Parallel Processing: System can support multiple actions executing simultaneously.
  • Terms often used interchangeably but are distinct concepts.

Key Question

• How does having multiple cores affect the speed of processing?

Conclusion

• Multi-core and parallel systems increase processing capability but are limited by software design and task nature.
• Amdahl's Law provides insight into potential speed increases.
• Understanding the difference between parallel and concurrent processing is crucial.