Improving Locality for Adaptive Irregular Scientific Codes

Abstract

An important class of scientific codes access memory in an irregular manner. Because irregular access patterns reduce temporal and spatial locality, they tend to underutilize caches, resulting in poor performance. Researchers have shown that consecutively packing data relative to traversal order can significantly reduce cache miss rates by increasing spatial locality. In this paper, we investigate techniques for using partitioning algorithms to improve locality in adaptive irregular codes. We develop parameters to guide both geometric (RCB) and graph partitioning (METIS) algorithms, and develop a new graph partitioning algorithm based on hierarchical clustering (GPART) which achieves good locality with low overhead. We also examine the effectiveness of locality optimizations for adaptive codes, where connection patterns dynamically change at intervals during program execution. We use a simple cost model to guide locality optimizations when access patterns change. Experiments on irregular scientific codes for a variety of meshes show our partitioning algorithms are effective for static and adaptive codes on both sequential and parallel machines. Improved locality also enhances the effectiveness of LOCALWRITE, a parallelization technique for irregular reductions based on the owner computes rule.

A Fast Parallel Graph Partitioner for Shared-Memory Inspector/Executor Strategies

“…To deal with this, the partitioner follows up the inflation step with a hierarchical partition joining step similar to GPart [7]. This phase is expressed in pseudocode in Figure 4.…”

“…GPart [7] is a hierarchical clustering partitioner that is significantly faster than METIS, but to the best of our knowledge it has not been parallelized. Its use has focused more on data reordering to improve locality [6] and it has features that allow the partitioning to target multiple levels of cache.…”

Intermediately executed code is the key to find refactorings that improve temporal data locality

“…On the other hand, temporal locality can only be improved by reordering the memory accesses so that the same addresses are accessed closer together. Advanced compiler methods to do this all target specific code patterns such as affine array expressions in regular loop nests [11,18,22], or specific sparse matrix computations [14,15,24,27]. For more general program constructs, fully-automatic optimization seems to be very hard, mainly due to the difficulty of the required dependence analysis.…”

A Hierarchical Model of Reference Affinity

“…Al-Furaih and Ranka examined graph-based clustering of irregular data for cache [3]. Other models include consecutive packing by Ding and Kennedy [10], space-filling curve by MellorCrummey et al [25], graph partitioning by Han and Tseng [17], and bucket sorting by Michell et al [26]. Several studies found that consecutive packing compared favorably with other models [25,31].…”