Sciweavers

Share
TPDS
2002

Automatic Partitioning of Parallel Loops with Parallelepiped-Shaped Tiles

10 years 11 months ago
Automatic Partitioning of Parallel Loops with Parallelepiped-Shaped Tiles
In this paper, an efficient algorithm to implement loop partitioning is introduced and evaluated. We start from results of Agarwal et al. [1] whose aim is to minimize the number of accessed data throughout the computation of a tile; this number is called the cumulative footprint of the tile. We improve these results along several directions. First, we derive a new formulation of the cumulative footprint, allowing for an analytical solution of the optimization problem stated in [1]. Second, we deal with arbitrary parallelepipedshaped tiles, as opposed to rectangular tiles in [1]. We design an efficient heuristic to determine the optimal tile shape in this general setting and we show its usefulness using both examples from [1] and a large collection of randomly generated data.
Fabrice Rastello, Yves Robert
Added 23 Dec 2010
Updated 23 Dec 2010
Type Journal
Year 2002
Where TPDS
Authors Fabrice Rastello, Yves Robert
Comments (0)
books