TY - GEN
T1 - Strategies to Improve the Performance of a Geophysics Model for Different Manycore Systems
AU - Serpa, Matheus S.
AU - Cruz, Eduardo H.M.
AU - Diener, Matthias
AU - Krause, Arthur M.
AU - Farres, Albert
AU - Rosas, Claudia
AU - Panetta, Jairo
AU - Hanzich, Mauricio
AU - Navaux, Philippe O.A.
N1 - Funding Information:
ACKNOWLEDGMENT This research received funding from the EU H2020 Programme and from MCTI/RNP-Brazil under the HPC4E project,grantn.o689772.ItwasalsosupportedbyIntelunder the Modern Code Project, and Petrobras.
Publisher Copyright:
© 2017 IEEE.
PY - 2017/11/14
Y1 - 2017/11/14
N2 - Many software mechanisms for geophysics exploration in Oil & Gas industries are based on wave propagation simulation. To perform such simulations, state-of-art HPC architectures are employed, generating results faster and with more accuracy at each generation. The software must evolve to support the new features of each design to keep performance scaling. Furthermore, it is important to understand the impact of each change applied to the software, in order to improve the performance as most as possible. In this paper, we propose several optimization strategies for a wave propagation model for five architectures: Intel Haswell, Intel Knights Corner, Intel Knights Landing, NVIDIA Kepler and NVIDIA Maxwell. We focus on improving the cache memory usage, vectorization, and locality in the memory hierarchy. We analyze the hardware impact of the optimizations, providing insights of how each strategy can improve the performance. The results show that NVIDIA Maxwell improves over Intel Haswell, Intel Knights Corner, Intel Knights Landing and NVIDIA Kepler performance by up to 17.9x.
AB - Many software mechanisms for geophysics exploration in Oil & Gas industries are based on wave propagation simulation. To perform such simulations, state-of-art HPC architectures are employed, generating results faster and with more accuracy at each generation. The software must evolve to support the new features of each design to keep performance scaling. Furthermore, it is important to understand the impact of each change applied to the software, in order to improve the performance as most as possible. In this paper, we propose several optimization strategies for a wave propagation model for five architectures: Intel Haswell, Intel Knights Corner, Intel Knights Landing, NVIDIA Kepler and NVIDIA Maxwell. We focus on improving the cache memory usage, vectorization, and locality in the memory hierarchy. We analyze the hardware impact of the optimizations, providing insights of how each strategy can improve the performance. The results show that NVIDIA Maxwell improves over Intel Haswell, Intel Knights Corner, Intel Knights Landing and NVIDIA Kepler performance by up to 17.9x.
KW - Geophysics
KW - Manycore systems
KW - Memory Hierarchy
KW - Vectorization
UR - http://www.scopus.com/inward/record.url?scp=85040541123&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85040541123&partnerID=8YFLogxK
U2 - 10.1109/SBAC-PADW.2017.17
DO - 10.1109/SBAC-PADW.2017.17
M3 - Conference contribution
AN - SCOPUS:85040541123
T3 - Proceedings - 29th International Symposium on Computer Architecture and High Performance Computing Workshops, SBAC-PADW 2017
SP - 49
EP - 54
BT - Proceedings - 29th International Symposium on Computer Architecture and High Performance Computing Workshops, SBAC-PADW 2017
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 29th International Symposium on Computer Architecture and High Performance Computing Workshops, SBAC-PADW 2017
Y2 - 17 October 2017 through 20 October 2017
ER -