Towards Hybrid Supercomputing Architectures
Abstract
In light of recent work on combining control-flow and dataflow architectures on the same chip die, a new architecture based on an asymmetric multicore processor is proposed.
The control-flow architectures are described as a most commonly used computer architecture today. Both multicore and manycore architectures are explained, as they are based on the same principles. A dataflow computing model assumes that data input flows through hardware as either a software or hardware dataflow implementation. In software dataflow, processors based on the control-flow paradigm process tasks based on their availability from the same queue (if there are any). In hardware dataflow architectures, the hardware is configured for a particular algorithm, and data input is streamed into the hardware, and the output is streamed back to the multicore processor for further processing. Hardware dataflow architectures are usually implemented with FPGAs.
Hybrid architectures employ asymmetric multicore and manycore computer architectures that are based on the control-flow and hardware dataflow architecture, all combined on the same chip die. Advantages include faster processing time, lower power consumption (and heating), and less space needed for the hardware.
References
[1] A. Kos, V. Ranković, and S. Tomažič, “Sorting networks on Maxeler dataflow supercomputing systems,” Advances in Computers, Vol. 96, pp. 139–186, 2015.
[2] V. Ranković, A. Kos, and V. Milutinović, “Bitonic merge sort implementation on the maxeler dataflow supercomputing system,” The IPSI BgD Transactions on Internet Research, Vol. 9, No. 2, pp. 5‒10, 2013.
[3] N. Korolija, V. Milutinovic, and S. Milosevic, “Accelerating conjugate gradient solver: temporal versus spatial data,” The IPSI BgD Transactions on Advanced Research, Vol. 3, No. 1, pp. 21–25, 2007.
[4] V. Milutinovic, M. Kotlar, S. Stojanovic, I. Dundic, N. Trifunovic, and Z. Babovic, “Implementing Neural Networks by Using the DataFlow Paradigm,” In DataFlow Supercomputing Essentials, New York: Springer Cham, 2017, pp. 3–44.
[5] V. Jelisavcic, I. Stojkovic, V. Milutinovic, and Z. Obradovic, “Fast learning of scale-free networks based on Cholesky factorization,” International Journal of Intelligent Systems, Vol. 33, No. 6, pp. 1322‒1339, 2018.
[6] M. J. Flynn, O. Mencer, V. Milutinovic, G. Rakocevic, P. Stenstrom, R. Trobec, and M. Valero, “Moving from petaflops to petadata,” Communications of the ACM, Vol. 56, No. 5, pp. 39‒42, 2013.
[7] A. Kos, S. Tomažič, J. Salom, N. Trifunovic, M. Valero, and V. Milutinovic, “New benchmarking methodology and programming model for big data processing,” International Journal of Distributed Sensor Networks, Vol. 11, No. 8, 271752, 2015. Available: https://journals.sagepub.com/doi/10.1155/2015/271752
[8] N. Trifunovic, V. Milutinovic, J. Salom, and A. Kos, “Paradigm shift in big data supercomputing: dataflow vs. controlflow,” Journal of Big Data, Vol. 2, No. 1, pp. 1‒9, 2015.
[9] D. G. Murray, F. McSherry, R. Isaacs, M. Isard, P. Barham, and M. Abadi, “Naiad: a timely dataflow system,” In Proceedings of the Twenty-Fourth ACM Symposium on Operating Systems Principles, 2013, pp. 439‒455.
[10] A. F. Gates, O. Natkovich, S. Chopra, P. Kamath, S. M. Narayanamurthy, C. Olston, ... and U. Srivastava, “Building a high-level dataflow system on top of Map-Reduce: the Pig experience,” Proceedings of the VLDB Endowment, Vol. 2, No. 2, pp. 1414‒1425, 2009.
[11] V. Milutinović, J. Salom, N. Trifunović, and R. Giorgi, Guide to dataflow supercomputing: Basic Concepts, Case Studies, and a Detailed Example, Springer Cham, 2015. Available: https://link.springer.com/book/10.1007/978-3-319-16229-4
[12] A. R. Hurson and V. Milutinovic, Dataflow Processing. Cambridge, Massachusetts, United States: Academic Press, 2015.
[13] V. Milutinović, B. Furht, Z. Obradović, and N. Korolija, “Advances in high performance computing and related issues,” Mathematical Problems in Engineering, Vol. 2016, 2016. Available: https://downloads.hindawi.com/journals/mpe/2016/2632306.pdf
[14] S. Stojanović, D. Bojić, and M. Bojović, “An overview of selected heterogeneous and reconfigurable architectures,” Advances in Computers, Vol. 96, pp. 1‒45, 2015.
[15] N. Korolija, T. Djukic, V. Milutinovic, and N. Filipovic, “Accelerating Lattice- Boltzmann method using Maxeler dataflow approach,” The IPSI BgD Transactions on Internet Research, Vol. 9, No. 2, pp. 34–42, 2013. Available: http://tir.ipsitransactions.org/indexTIR_all.php
[16] S. Stojanović, D. Bojić, and V. Milutinović, “Solving Gross Pitaevskii equation using dataflow paradigm,” The IPSI BgD Transactions on Internet Research, Vol. 17, 2013. Available: http://ipsitransactions.org/journals/papers/tir/2013july/p4.pdf
[17] I. Stanojević, M. Kovačević, and V. Šenk, “Application of maxeler dataflow supercomputing to spherical code design,” In Exploring the DataFlow Supercomputing Paradigm, New York: Springer Cham, 2019, pp. 133‒168.
[18] N. Bežanić, J. Popović-Božović, V. Milutinović, and I. Popović, “Implementation of the RSA Algorithm on a DataFlow Architecture,” IPSI BgD Transactions on Internet Research, Vol. 9, No. 2, pp. 11‒16, 2013.
[19] N. Korolija, J. Popović, M. Cvetanović, and M. Bojović, “Dataflow-based parallelization of control-flow algorithms,” Advances in Computers, Vol. 104, pp. 73‒124, 2017.
[20] V. Milutinovic, J. Salom, D. Veljovic, N. Korolija, D. Markovic, and L. Petrovic, “Transforming applications from the control flow to the dataflow paradigm,” In Dataflow Supercomputing Essentials, New York: Springer Cham, 2017, pp. 107‒129.
[21] J. Popovic, D. Bojic, and N. Korolija, “Analysis of task effort estimation accuracy based on use case point size,” IET Software, Vol. 9, No. 6, pp. 166‒173, 2015.
[22] N. Trifunovic, V. Milutinovic, N. Korolija, and G. Gaydadjiev, “An AppGallery for dataflow computing,” Journal of Big Data, Vol. 3, No. 1, pp. 1‒30, 2016.
[23] V. Milutinovic, J. Salom, D. Veljovic, N. Korolija, D. Markovic, and L. Petrovic, “Maxeler AppGallery Revisited,” In Dataflow Supercomputing Essentials, New York: Springer Cham, 2017, pp. 3‒18.
[24] A. Hurson, V. Milutinovic, “Special issue on dataflow supercomputing,” Advances in Computers, Vol. 96, pp. 1–234, 2015.
[25] N. Korolija, D. Bojic, A. R. Hurson, and V. Milutinovic, “A runtime job scheduling algorithm for cluster architectures with dataflow accelerators,” Advances in Computers, Vol. 126, pp. 201-245, 2022.
[26] V. Milutinović, E. S. Azer, K. Yoshimoto, G. Klimeck, M. Djordjevic, M. Kotlar, ... and I. Ratkovic, “The ultimate dataflow for ultimate supercomputers-on-a-chip, for scientific computing, geo physics, complex mathematics, and information processing,” In 2021 10th Mediterranean Conference on Embedded Computing (MECO), 2021, pp. 1‒6.
[27] V. Milutinović, N. Trifunović, N. Korolija, J. Popović, and D. Bojić, “Accelerating program execution using hybrid control flow and dataflow architectures,” In 2017 25th Telecommunication Forum (TELFOR), 2017, pp. 1‒4.
[28] Z. Sustran and J. Protic, “Migration in hardware transactional memory on asymmetric multiprocessor,” IEEE Access, Vol. 9, pp. 69346‒69364, 2021.
[29] J. Popović, V. Jelisavčić, and N. Korolija, “Hybrid Supercomputing Architectures for Artificial Intelligence: Analysis of Potentials,” In 1st Serbian International Conference on Applied Artificial Intelligence (SICAAI), Kragujevac, Serbia, 2022.
[30] N. Korolija and A. Zamuda, “On Cloud-Supported Web-Based Integrated Development Environment for Programming DataFlow Architectures,” In Exploring the DataFlow Supercomputing Paradigm, New York: Springer Cham, 2019, pp. 41‒51
[31] K. Huang, Y. Liu, N. Korolija, J. M. Carulli, and Y. Makris, “Recycled IC detection based on statistical methods,” IEEE transactions on computer-aided design of integrated circuits and systems, Vol. 34, No. 6, pp. 947‒960, 2015.
I (we), the author(s), hereby declare under full moral, financial and criminal liability that the manuscript submitted for publication to the Journal of Computer and Forensic Sciences
a) is the result of my (our) own original research and that I (we) hold the right to publish it;
b) does not infringe any copyright or other third-party proprietary rights;
c) complies with the Journal’s research and publishing ethics standards;
d) has not been published elsewhere, under this or any other title;
e) is not under consideration by another publication, under this or any other title.
I (we) also declare under full moral, financial and criminal liability:
f) that all conflicts of interest that may directly or potentially influence or impart bias on the work have been disclosed in the manuscript;
g) that if the article has been accepted for publishing I (we) will transfer all copyright ownership of the manuscript to the University of Criminal Investigation and Police Studies in Belgrade.
Signed by the Corresponding Author on behalf of the all other authors.