Abstract
Various existing optimization and memory consistency management techniques for GPU applications rely on memory access patterns of kernels. However, they suffer from poor practicality because they require explicit user interventions to extract kernel memory access patterns. This paper proposes an automatic memory-access-pattern analysis framework called MAPA. MAPA is based on a source-level analysis technique derived from traditional symbolic analyses and a run-time pattern selection technique. The experimental results show that MAPA properly analyzes 116 real-world OpenCL kernels from Rodinia and Parboil.
- R. A. Ballance, A. B. Maccabe, and K. J. Ottenstein. The program dependence web: A representation supporting control-, data-, and demand-driven interpretation of imperative languages. In phProceedings of the ACM SIGPLAN 1990 Conference on Programming Language Design and Implementation, pages 257--271, 1990. Google Scholar
Digital Library
- M. Bauer, H. Cook, and B. Khailany. CudaDMA: Optimizing GPU memory bandwidth via warp specialization. In phProceedings of 2011 International Conference for High Performance Computing, Networking, Storage and Analysis, 2011. Google Scholar
Digital Library
- S. Che, J. W. Sheaffer, and K. Skadron. Dymaxion: Optimizing memory access patterns for heterogeneous systems. In phProceedings of 2011 International Conference for High Performance Computing, Networking, Storage and Analysis, 2011. Google Scholar
Digital Library
- M. R. Haghighat and C. D. Polychronopoulos. Symbolic analysis for parallelizing compilers. In phACM Transactions on Programming Languages and Systems, volume 18, pages 477--518, 1996. Google Scholar
Digital Library
- B.-N. Tal, E. Levy, A. Barak, and E. Rubin. Memory access patterns: The missing piece of the multi-GPU puzzle. In phProceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, 2015. Google Scholar
Digital Library
- P. Tu and D. Padua. Gated SSA-based demand-driven symbolic analysis for parallelizing compilers. In phProceedings of the 9th International Conference on Supercomputing, pages 414--423, 1995. Google Scholar
Digital Library
Index Terms
POSTER: MAPA: An Automatic Memory Access Pattern Analyzer for GPU Applications
Recommendations
POSTER: MAPA: An Automatic Memory Access Pattern Analyzer for GPU Applications
PPoPP '17: Proceedings of the 22nd ACM SIGPLAN Symposium on Principles and Practice of Parallel ProgrammingVarious existing optimization and memory consistency management techniques for GPU applications rely on memory access patterns of kernels. However, they suffer from poor practicality because they require explicit user interventions to extract kernel ...
Accelerating financial applications on the GPU
GPGPU-6: Proceedings of the 6th Workshop on General Purpose Processor Using Graphics Processing UnitsThe QuantLib library is a popular library used for many areas of computational finance. In this work, the parallel processing power of the GPU is used to accelerate QuantLib financial applications. Black-Scholes, Monte-Carlo, Bonds, and Repo code paths ...
Multi-kernel Auto-Tuning on GPUs: Performance and Energy-Aware Optimization
PDP '15: Proceedings of the 2015 23rd Euromicro International Conference on Parallel, Distributed, and Network-Based ProcessingPrompted by their very high computational capabilities and memory bandwidth, Graphics Processing Units (GPUs) are already widely used to accelerate the execution of many scientific applications. However, programmers are still required to have a very ...







Comments