Abstract
Recently, it has been common to use parallel processing for machine learning. CGRAs are drawing attention in terms of reconfigurability and high performance. We propose a method to map data-flow graphs onto CGRAs by SAT solving. The proposed method can perform the automatic transformation which changes the order of operations in data-flow graphs to obtain more efficient schedules. It also accommodates mapping of multi-node operations like MAC operation. We have solved mapping problems of matrix-vector multiplication. In our experiment, a SAT solver outperformed an ILP solver. Our method successfully processed a data-flow graph of more than a hundred nodes. The automatic transformation under the associative and commutative laws was not as much scalable but successfully reduced the number of cycles, where the XBTree-based method worked faster than the enumeration-based method. As another direction, we tried to optimize a CGRA architecture according to a data-flow graph and were able to reduce its PEs and connections through incremental SAT solving.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Krizhevsky, A., Sutskever, I., Hinton, G.E.: ImageNet classification with deep convolutional neural networks. In: Proceedings of International Conference on Neural Information Processing Systems, pp. 1097–1105 (2012)
Jouppi, N.P., et al.: In-datacenter performance analysis of a tensor processing unit. ACM SIGARCH Comput. Archit. News 45(2), 1–12 (2017). https://doi.org/10.1145/3140659.3080246
Liu, L., et al.: A survey of coarse-grained reconfigurable architecture and design. ACM Comput. Surv. (CSUR) 52(6), 1–39 (2020). https://doi.org/10.1145/3357375
Yoshida, H., Fujita, M.: Exact minimum factoring of incompletely specified logic functions via quantified Boolean satisfiability. IPSJ Trans. Syst. LSI Des. Methodol. 4, 70–79 (2011). https://doi.org/10.2197/ipsjtsldm.4.70
Audemard, G., Lagniez, J.-M., Simon, L.: Improving glucose for incremental SAT solving with assumptions: application to MUS extraction. In: Järvisalo, M., Van Gelder, A. (eds.) SAT 2013. LNCS, vol. 7962, pp. 309–317. Springer, Heidelberg (2013). https://doi.org/10.1007/978-3-642-39071-5_23
Mei, B., Vernalde, S., Verkest, D., De Man, H., Lauwereins, R.: Exploiting loop-level parallelism on coarse-grained reconfigurable architectures using modulo scheduling. IEE Proc. Comput. Digit. Tech. 150(5), 255 (2003). https://doi.org/10.1049/ip-cdt:20030833
Chin, S.A., Anderson, J.H.: An architecture-agnostic integer linear programming approach to CGRA mapping. In: Proceedings of Design Automation Conference (DAC), pp. 1–6 (2018). https://doi.org/10.1145/3195970.3195986
Greene, J.W.: Exact mapping of rewritten linear functions to configurable logic. In: Proceedings of International Workshop on FPGAs for Software Programmers (FSP), pp. 11–18 (2019)
Chin, S.A., et al.: CGRA-ME: a unified framework for CGRA modelling and exploration. In: Proceedings of International Conference on Application-specific Systems, Architectures and Processors (ASAP), pp. 184–189 (2017). https://doi.org/10.1109/ASAP.2017.7995277
Flynn, M.J., Pell, O., Mencer, O.: Dataflow supercomputing. In: Proceedings of International Conference on Field Programmable Logic and Applications (FPL), pp. 1–3 (2012). https://doi.org/10.1109/FPL.2012.6339170
Miyasaka, Y., Fujita, M.: SAT-based mapping of data-flow onto array processor. In: 2020 IFIP/IEEE International Conference on Very Large Scale Integration (VLSI-SoC) (2020)
Nguyen, V.H., Mai, S.T.: A new method to encode the at-most-one constraint into SAT. In: Proceedings of International Symposium on Information and Communication Technology (SoICT), 03–04 December, pp. 1–8 (2015). https://doi.org/10.1145/2833258.2833293
Sinz, C.: Towards an optimal CNF encoding of Boolean cardinality constraints. In: van Beek, P. (ed.) CP 2005. LNCS, vol. 3709, pp. 827–831. Springer, Heidelberg (2005). https://doi.org/10.1007/11564751_73
Lee, G., Choi, K., Dutt, N.D.: Mapping multi-domain applications onto coarse-grained reconfigurable architectures. IEEE Trans. Comput. Aided Design Integr. Circuits Syst. 30(5), 637–650 (2011). https://doi.org/10.1109/TCAD.2010.2098571
Yoon, J., Shrivastava, A., Park, S., Ahn, M., Paek, Y.: A graph drawing based spatial mapping algorithm for coarse-grained reconfigurable architectures. IEEE Trans. Very Large Scale Integr. (VLSI) Syst. 17(11), 1565–1578 (2009). https://doi.org/10.1109/TVLSI.2008.2001746
Živković, M.: Classification of small (0,1) matrices. Linear Algebra Appl. 414(1), 310–346 (2006). https://doi.org/10.1016/j.laa.2005.10.010
Liu, B., Baas, B.M.: Parallel AES encryption engines for many-core processor arrays. IEEE Trans. Comput. 62(3), 536–547 (2013). https://doi.org/10.1109/TC.2011.251
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2021 IFIP International Federation for Information Processing
About this paper
Cite this paper
Miyasaka, Y., Fujita, M., Mishchenko, A., Wawrzynek, J. (2021). SAT-Based Mapping of Data-Flow Graphs onto Coarse-Grained Reconfigurable Arrays. In: Calimera, A., Gaillardon, PE., Korgaonkar, K., Kvatinsky, S., Reis, R. (eds) VLSI-SoC: Design Trends. VLSI-SoC 2020. IFIP Advances in Information and Communication Technology, vol 621. Springer, Cham. https://doi.org/10.1007/978-3-030-81641-4_6
Download citation
DOI: https://doi.org/10.1007/978-3-030-81641-4_6
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-81640-7
Online ISBN: 978-3-030-81641-4
eBook Packages: Computer ScienceComputer Science (R0)