Explicit parallelism: Difference between revisions

Revision as of 12:59, 4 February 2024

Latest revision as of 00:40, 4 August 2025

(One intermediate revision by one other user not shown)

Line 3:

{{Use mdy dates |date=February 2024}}

{{More citations needed |date=February 2024}}

In [[computer programming]], '''explicit parallelism''' is the representation of concurrent computations using primitives in the form of operators, function calls or special-purpose directives.<ref name="pra11" /> Most parallel primitives are related to process synchronization, communication and process partitioning.<ref name="dij68" /> As they seldomrarely contribute to actually carry out the intended computation of the program but, rather, structure it, their computational cost is often considered as overhead.

The advantage of explicit [[parallel programming]] is increased programmer control over the computation. A skilled parallel programmer may take advantage of explicit parallelism to produce efficient code for a given target computation environment. However, programming with explicit parallelism is often difficult, especially for non-computing specialists, because of the extra work and skill involved in developing it.

The advantage of explicit [[parallel programming]] is increased programmer

control over the computation. A skilled parallel programmer may take advantage of explicit parallelism to produce efficient code for a given target computation environment. However, programming with explicit parallelism is often difficult, especially for non-computing specialists, because of the extra work and skill involved in developing it.

In some instances, explicit parallelism may be avoided with the use of an optimizing compiler or runtime that automatically deduces the parallelism inherent to computations, known as [[implicit parallelism]].

Line 24 ⟶ 23:

{{reflist |refs=

<ref name="dij68">{{cite journal |title=The structure of the "THE"-multiprogramming system|date=1968-05-01 |first=Edsger W. |last=Dijkstra |author-link=Edsger W. Dijkstra |journal=Communications of the ACM |volume=11 |issue=5 |pages=341–346 |doi=10.1145/363095.363143 }}</ref>

<ref name="pra11">{{cite conference |title=Parallel programming: design of an overview class |date=June 2011 |first=Christoph |last=von Praun |conference=Proceedings of the 2011 ACM SIGPLAN X10 Workshop |number=2 |pages=1–6 |doi=10.1145/2212736.2212738 }}</ref>

}}