Memory access pattern: Difference between revisions

Revision as of 01:09, 14 June 2016

Revision as of 01:10, 14 June 2016

Line 37:

=== Nearest neighbour ===

Nearest neighbour memory access patterns appear in simulation, and are related to sequential/strided patterns. An algorithm may traverse a data structure using information from the nearest neighbours of a data element (in one or more dimensions) to perform a calculation. These are common in physics simulations operating on grids.<ref name="PGAS programming"/>. Nearest neighbour can also refer to inter-node communication in a cluster; Physics simulations which rely on such local access patterns can be paralleledparallelized andwith mappedthe ontodata apartitioned into cluster nodes, with purely nearest-neighbour communication between them, which may have advantages for latency and communication bandwidth. This use case maps well onto [[torus network topology]].

<ref>{{cite web|title=Quantifying Locality In The Memory Access Patterns of HPC Applications|url=http://www.sdsc.edu/~allans/sc05_locality.pdf}}mentions nearest neighbour access patterns in clusters</ref>