Concurrency - Parallel Execution

Introduction

Where a sequential execution can be imagined as a single worker on an assembly line moving between tasks, a parallel process is more like a series of workers, each doing a specific task.

Parallelism use is only going to increase because:

• CPU clock speeds are barely increasing (Moores law hits the roof),
• Multi-core processors are standard,
• and networks are getting faster.

Parallel execution is the ability to :

• divide a single operation through the hardware ressource
• apply multiple CPU and I/O resources to the execution of a single operation.

There are three architectures and approaches to increase performance through paralleled work over additional hardware:

• shared disk known also as shared memory or shared everything
• shared nothing

See also: Parallel Programming - (Function|Operation)

Articles Related

Implementation

mechanism by which a task can be run in parallel may be:

• time-sharing cooperative multitasking,
• fibers,
• threads,
• processes, using different hyperthreads,
• cores,
• CPUs,
• machines,
• etc.

Documentation / Reference

• How to Build a High-Performance Data Warehouse By David J. DeWitt, Ph.D.; Samuel Madden, Ph.D.; and Michael Stonebraker, Ph.D.