TY - GEN
T1 - Performance Estimation for Exascale Reconfigurable Dataflow Platforms
AU - Yasudo, Ryota
AU - Coutinho, Jose
AU - Varbanescu, Ana
AU - Luk, Wayne
AU - Amano, Hideharu
AU - Becker, Tobias
N1 - Funding Information:
Acknowledgments. This work was supported by the EU H2020 Research and Innovation Programme under grant agreement number 671653 and JST/CREST program entitled “Research and Development on Unified Environment of Accelerated Computing and Interconnection for Post-Petascale Era” in the research area of “Development of System Software Technologies for post-Peta Scale High Performance Computing”.
PY - 2018/12
Y1 - 2018/12
N2 - The next generation high-performance computing platforms will need to support exascale computing. A promising path in achieving exascale is to embrace heterogeneity and specialised computing in the form of reconfigurable accelerators. However, assessing the feasibility of heterogeneous exascale systems requires fast and accurate performance prediction. This paper proposes PERKS, a novel performance estimation frame-work for reconfigurable dataflow platforms (RDPs). PERKS uses machine and application parameters to build an analytical model for predicting the performance of multi-Accelerator systems. Moreover, model calibration is automatic, making the model flexible and usable for different machine configurations and applications. Our experimental results demonstrate that PERKS can predict the performance of current workloads and RDPs with an accuracy above 95%. We also demonstrate how the modelling scales to exascale workloads and exascale platforms.
AB - The next generation high-performance computing platforms will need to support exascale computing. A promising path in achieving exascale is to embrace heterogeneity and specialised computing in the form of reconfigurable accelerators. However, assessing the feasibility of heterogeneous exascale systems requires fast and accurate performance prediction. This paper proposes PERKS, a novel performance estimation frame-work for reconfigurable dataflow platforms (RDPs). PERKS uses machine and application parameters to build an analytical model for predicting the performance of multi-Accelerator systems. Moreover, model calibration is automatic, making the model flexible and usable for different machine configurations and applications. Our experimental results demonstrate that PERKS can predict the performance of current workloads and RDPs with an accuracy above 95%. We also demonstrate how the modelling scales to exascale workloads and exascale platforms.
KW - FPGAs
KW - Performance modelling
KW - exascale computing
KW - heterogeneous systems
KW - reconfigurable platforms
UR - http://www.scopus.com/inward/record.url?scp=85068319351&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85068319351&partnerID=8YFLogxK
U2 - 10.1109/FPT.2018.00062
DO - 10.1109/FPT.2018.00062
M3 - Conference contribution
AN - SCOPUS:85068319351
T3 - Proceedings - 2018 International Conference on Field-Programmable Technology, FPT 2018
SP - 317
EP - 320
BT - Proceedings - 2018 International Conference on Field-Programmable Technology, FPT 2018
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 17th International Conference on Field-Programmable Technology, FPT 2018
Y2 - 10 December 2018 through 14 December 2018
ER -