ADHA: Automatic Data layout framework for Heterogeneous Architectures
Rice University
arXiv:1407.4859 [cs.DC], (18 Jul 2014)
@article{majeti2014adha,
title={ADHA: Automatic Data layout framework for Heterogeneous Architectures},
author={Majeti, Deepak and Meel, Kuldeep S and Barik, Rajkishore and Sarkar, Vivek},
journal={arXiv preprint arXiv:1407.4859},
year={2014}
}
Data layouts play a crucial role in determining the performance of a given application running on a given architecture. Existing parallel programming frameworks for both multicore and heterogeneous systems leave the onus of selecting a data layout to the programmer. Therefore, shifting the burden of data layout selection to optimizing compilers can greatly enhance programmer productivity and application performance. In this work, we introduce {ADHA}: a two-level hierarchal formulation of the data layout problem for modern heterogeneous architectures. We have created a reference implementation of ADHA in the Heterogeneous Habanero-C (H2C) parallel programming system. ADHA shows significant performance benefits of up to 6.92x compared to manually specified layouts for two benchmark programs running on a CPU+GPU heterogeneous platform.
July 24, 2014 by hgpu
