Name
Abstract | ||
|---|---|---|
Data transformation is one of the key optimizations in maximizing cache locality. Traditional data transformation strategies employ linear data layouts, e. g., row-major or column-major, for multidimensional arrays. Although a linear layout matches the linear memory space well in most cases, it can only optimize for self-spatial locality for individual references. In this work, we propose a novel data layout transformation framework that is able to determine a tiled layout for each array in an application program. Tiled layout can exploit the group-spatial locality among different references and improve cache line utilization. In our strategy, the data elements accessed by different references in one loop iteration are placed into a tile and fetched into the same cache line at runtime. This helps minimizing conflict misses in caches. We evaluated our data layout transformation framework using 30 benchmarks on a commercial multicore machine. The experimental results show that our approach outperforms state-of-the-art data transformation strategies and works well with large core counts.
|
Year | DOI | Venue |
|---|---|---|
2011 | 10.1109/ICCAD.2011.6105318 | ICCAD |
Keywords | Field | DocType |
linear data layout,commercial multicore machine,optimizing data,data element,cache storage,array tiling,cache locality maximization,traditional data transformation strategy,tiled layout,linear layout,state-of-the-art data transformation strategy,linear data layouts,row-major,column-major,data layout transformation framework,data transformation,multiprocessing systems,different reference,selfspatial locality,novel data layout transformation,data locality optimization,multidimensional arrays,group-spatial locality,timing analysis,thermal | Locality,Data layout,Locality of reference,Cache,CPU cache,Computer science,Parallel computing,Exploit,Static timing analysis,Multi-core processor | Conference |
ISSN | ISBN | Citations |
1092-3152 E-ISBN : 978-1-4577-1398-9 | 978-1-4577-1398-9 | 1 |
PageRank | References | Authors |
0.36 | 24 | 4 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Wei Ding | 1 | 130 | 11.67 |
| Yuanrui Zhang | 2 | 180 | 15.48 |
| Jun Liu | 3 | 48 | 4.44 |
| Mahmut T. Kandemir | 4 | 7371 | 568.54 |