Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
PNNL: High Performance Computing

Global Arrays Programming Models

Global Arrays (GA) is a Partitioned Global Address Space (PGAS) programming model. It provides primitives for one-sided communication (Get, Put, Accumulate) and Atomic Operations (read increment). It supports blocking and non-blocking primtives, and supports location consistency.

Starting with GA v5.4, ARMCI has been deprecated and replaced with Communication runtime for Extreme Scale (ComEx). GA uses ComEx to abstract inter-node communication operations. The default ComEx ports use MPI --- which makes GA and ComEx portable for high-end systems. Most ComEx implementations use on-node shared memory for faster intra-node communication.

Latest Release

- Release Notes - Download

Citing GA

Please refer to this section of our publications page.

Global Arrays Links

News and Events

  • November 2, 2020
    GA 5.8 released
  • March 30, 2018
    GA 5.7 released
  • March 29, 2018
    GA 5.6.5 released
  • May 30, 2017
    GA 5.6.1 released
  • April 4, 2017
    GA 5.6 released
  • August 3, 2016
    GA 5.5 released
  • April 12, 2016
    GA 5.4 released

Feedback

Contacts