Chapter 5 Parallel algorithms and patterns

This is from Chapter 5 of Parallel and High Performance Computing, Robey and Zamora, Manning Publications, available at http://manning.com

The book may be obtained at http://www.manning.com/?a_aid=ParallelComputingRobey

Copyright 2019-2020 Robert Robey, Yuliana Zamora, and Manning Publications Emails: [email protected], [email protected]

See License.txt for licensing information.

Particle collisions psuedocode (Book: listing 5.1)

PerfectHash hash table write and read (Book: listing 5.2 to 5.3)

 Used in following examples
 Uses autoconf and make
 Has OpenCL GPU code
 Configure ./configure
 Build with make

PerfectHash Neighbor finding (Book: listing 5.4 to 5.5)

 Run ./neigh and ./neigh2d
 Performance output for various sizes and compute hardware

PerfectHash Remap (Book: listing 5.6 to 5.7)

 Run ./remap and ./remap2d
 Performance output for various sizes and compute hardware

PerfectHash Table interpolation (Book: listing 5.8)

 Run ./table and ./tablelarge
 Performance output for various sizes and compute hardware

PerfectHash Sort (Book: listing 5.9)

 Run ./sort and ./sort
 Performance output for various sizes and compute hardware

CompactHash (Book: section 5.4.2)

 Build with cmake
 Configure with cmake
 Build with make
 Run ./neigh and ./neigh2d

CompactHashRemap (Book: listings 5.10 through 5.12)

 Build with cmake
 Configure with cmake
 Build with make
 Run AMR_remap/AMR_remap

PrefixSums (Book: pseudocode -- will have implementations in Chapters 6 & 7)

GlobalSums (Book: listings 5.15 through 5.21)

 Build with cmake
 Configure with cmake
 Build with make
 Run ./globalsums
 Outputs performance results for various sizes, hardware, and methods