Data Structures and Algorithms

   

Parallelisation with a Grid Abstraction

Authors: Sven de Smet

This paper describes a new technique for automatic parallelisation in the Z-polyhedral model. The presented technique is applicable to arbitrarily nested loopnests with iteration spaces that can be represented as unions of Z-polyhedra and affine modular data-access functions. The technique partitions both iteration and data spaces of the computation. The maximal amount of parallelism that can be represented using grid partitions is extracted.

Comments: 16 pages, This paper is a slightly modified version of a draft paper that was submitted to ParCo 2011 (with added proofs) and is very preliminary. Since I do not have the resources to complete this paper by increasing its clarity, adding examples, adding an experimental evaluation and adding a section on related work, I'm making it available so that it may be useful to others.

Download: PDF

Submission history

[v1] 12 Jun 2011

Unique-IP document downloads: 129 times

Add your own feedback and questions here:
You are equally welcome to be positive or negative about any paper but please be polite. If you are being critical you must mention at least one specific error, otherwise your comment will be deleted as unhelpful.

comments powered by Disqus