Determining the idle time of a tiling

Abstract

This paper investigates the idle time associated with a parallel computation, that is, the time that processors are idle because they are either waiting for data from other processors or waiting to synchronize with other processors. We study doubly-nested loops corresponding to parallelogram- or trapezoidal-shaped iteration spaces that have been parallelized by the well-known tiling transformation. We introduce the notion of rise r, which relates the shape of the iteration space to that of the tiles. For parallelogram- shaped iteration spaces, we show that when r < -2, the idle time is linear in P, the number of processors, but when r > -1, it is quadratic in P. In the context of hierarchical tiling, where multiple levels of tiling are used, a good choice of rise can lead to less idle time and better performance. While idle time is not the only cost that should be considered in evaluating a tiling strategy, current architectural trends (of deeper memory hierarchies and multiple levels of parallelism) suggest it has increasing importance.