Talk:Massively parallel (computing)
|This is the talk page for discussing improvements to the Massively parallel (computing) article.
This is not a forum for general discussion of the article's subject.
|WikiProject Computing||(Rated Start-class, Mid-importance)|
More specific definitions
The definition at the start of article that "massively parallel refers to the use of a large number of processors (or separate computers) to perform a set of coordinated computations in parallel (simultaneously)" is good. But the article does not warn that sometimes people have used the term MPP with various more specific definitions.
- "As Bell and Gray make clear, the term MPP, although used pervasively throughout the history of the Top500 list, is confusing, misleading, and provides little specification of system type. The notion of it has been abused, confused, and ironically derived from a different type of system than that to which it is ordinarily applied. (The original MPP was an SIMD-class computer, which is a separate classification of the Top500 list.)" [page 55] "MPP ... could mean a distributed-memory system, a large shared-memory system with or without cache coherence, a large vector system, and so on: it covers too many categories and hides too many salient differences to be a useful tool for description. On this point, we have come to agree with Bell and Gray." [page 57] Jack Dongarra et al, "High-performance computing: clusters, constellations, MPPs, and future directions", Computing in Science and Engineering, 7(2), 2005, 51-59.
- "Clusters with over 1000 processors were called massively parallel processors or MPPs." [page 92] Gordon Bell and Jim Gray, "What’s next in High-Performance Computing", Comm. ACM, 45(2), 2002, 91–95.
- Wikipedia at parallel computing says "MPPs have many of the same characteristics as clusters, but MPPs have specialized interconnect networks (whereas clusters use commodity hardware for networking)." And a Stack Overflow comment at What is the different between a Cluster and MPP supercomputer architecture agrees with this.
- According to TOP500, and as noted in this article, "MPP architectures are the second most common supercomputer implementations after clusters." But it is not clear what definitions they use: the definition of MPP cannot be just "a large number of processors" because all modern supercomputers satisfy that definition.