Next week is your chance to check out PGAS 2011 – the Fifth Conference on Partitioned Global Address Space Programing Models, which is taking place at Galveston Island, Texas on Oct. 15-18.
Partitioned Global Address Space (PGAS) programming models offer HPC programmers a shared address space model that simplifies programming while exposing data/thread locality to enhance performance. This facilitates the development of programming models that can deliver both productivity and performance. The PGAS conference is a forum to present and discuss ideas and research developments in the area of PGAS models, languages, compilers, runtimes, applications and tools.
In case you didn’t know, the PGAS parallel programming model assumes a global memory address space that is logically partitioned and a portion of it is local to each processor. According to the Wikipedia, the novelty of PGAS is that the portions of the shared memory space may have an affinity for a particular thread, thereby exploiting locality of reference. The PGAS model is the basis of Unified Parallel C, Co-array Fortran, Titanium, Fortress, Chapel and X10.
In related news, the Parallel Architectures and Compilation Techniques PACT Conference is taking place this week at the same location. The purpose of PACT is to bring together researchers from architecture, compilers, applications and languages to present and discuss innovative research of common interest.