Where Should The Threads Go? Leveraging Hierarchical Data Locality to Solve the Thread Affinity Dilemma

Ahmad Anbar, Abdel-Hameed Badawy, Olivier Serres, and Tarek El-Ghazawi (2014). Where Should The Threads Go? Leveraging Hierarchical Data Locality to Solve the Thread Affinity Dilemma. In: The 20th IEEE International Conference on Parallel and Distributed Systems (ICPADS 2014), December 16 – 19, 2014, Hsinchu, Taiwan. (In Press)

[img]
Preview
Text
paper.pdf

Download (490kB) | Preview

Abstract

We are proposing a novel framework that ameliorates locality-aware parallel programming models, by defining a hierarchical data locality model extension. We also propose two hierarchical thread partitioning algorithms. These algorithms synthesize hierarchical thread placement layouts that targets minimizing the program's overall communication costs. We demonstrate the effectiveness of our approach using the NAS Parallel Benchmarks implemented in Unified Parallel C (UPC) language using a modified Berkeley UPC Compiler and runtime system. We achieved performance gains of up to 88% in performance from applying the placement layouts our algorithms suggest.

Item Type: Conference or Workshop Item (Paper)
Uncontrolled Keywords: Data locality, Many-cores, Hierarchical thread clustering
Subjects: 01) General HPC
02) PGAS
06) Computer Architecture
07) Scheduling and Optimization
08) Design > 08.2) Software
Depositing User: Dr. Abdel-Hameed Badawy
Date Deposited: 14 Apr 2015 18:57
Last Modified: 14 Apr 2015 18:57
URI: http://hpcl.seas.gwu.edu:9000/eprints/id/eprint/38

Actions (login required)

View Item View Item