Abstract
As computing platforms increasingly embrace heterogeneity, runtime resource managers need to efficiently, dynamically, and robustly manage shared resources (e.g., cores, power budgets, memory bandwidth). To address the complexities in heterogeneous systems, state-of-the-art techniques that use heuristics or machine learning have been proposed. On the other hand, conventional control theory can be used for formal guarantees, but may face unmanageable complexity for modeling system dynamics of complex heterogeneous systems. We address this challenge through HESSLE-FREE (Heterogeneous Systems Leveraging Fuzzy Control for Runtime Resource Management): an approach leveraging fuzzy control theory that combines the strengths of classical control theory together with heuristics to form a light-weight, agile, and efficient runtime resource manager for heterogeneous systems. We demonstrate the efficacy of HESSLE-FREE executing on a NVIDIA Jetson TX2 platform (containing a heterogeneous multi-processor with a GPU) to show that HESSLE-FREE: 1) provides opportunity for optimization in the controller and stability analysis to enhance the confidence in the reliability of the system; 2) coordinates heterogeneous compute units to achieve desired objectives (e.g., QoS, optimal power references, FPS) efficiently and with lower complexity, and 3) eases the burden of system specification.
- A. M. Rahmani et al. 2015. Dynamic power management for many-core platforms in the dark silicon era: A multi-objective control approach. In ISLPED.Google Scholar
- J. Aracil and F. Gordillo. 2000. Stability Issues in Fuzzy Control. Physica-Verlag HD. https://books.google.com/books?id=h6weAQAAIAAJ.Google Scholar
- ARM. 2013. big.LITTLE Technology: The Future of Mobile. Technical Report. https://www.arm.com/files/pdf/big_LITTLE_Technology_the_Futue_of_Mobile.pdf.Google Scholar
- Andrea Bartolini, Matteo Cacciari, Andrea Tilli, and Luca Benini. [2011]. A distributed and self-calibrating model-predictive controller for energy and thermal management of high-performance multicores. In DATE, 2011.Google Scholar
- Christian Bienia, Sanjeev Kumar, Jaswinder Pal Singh, and Kai Li. 2008. The PARSEC benchmark suite. In PACT. ACM Press, New York, New York, USA, 72. DOI:https://doi.org/10.1145/1454115.1454128Google Scholar
- Ramazan Bitirgen et al. 2008. Coordinated management of multiple interacting resources in chip multiprocessors: A machine learning approach. In MICRO.Google Scholar
- R. Bitirgen, E. Ipek, and J. F. Martinez. 2008. Coordinated management of multiple interacting resources in chip multiprocessors: A machine learning approach. In 2008 41st IEEE/ACM International Symposium on Microarchitecture.Google Scholar
- Tobias Bjerregaard and Shankar Mahadevan. 2006. A survey of research and practices of network-on-chip. Comput. Surveys 38, 1 (2006).Google Scholar
- Zhijia Chen, Yuanchang Zhu, Yanqiang Di, and Shaochong Feng. 2015. A dynamic resource scheduling method based on fuzzy control theory in cloud environment. Journal of Control Science and Engineering 2015 (2015), 1--10. DOI:https://doi.org/10.1155/2015/383209Google Scholar
Digital Library
- NVIDIA Corporation. 2017. NVIDIA Jetson TX2 embedded module. https://developer.nvidia.com/embedded/buy/jetson-tx2.Google Scholar
- L. Costero, A. Iranfar, M. Zapater, F. D. Igual, K. Olcoz, and D. Atienza. 2019. MAMUT: Multi-agent reinforcement learning for efficient real-time multi-user video transcoding. In 2019 Design, Automation Test in Europe Conference Exhibition (DATE).Google Scholar
- Christina Delimitrou and Christos Kozyrakis. 2014. Quasar: Resource-efficient and QoS-aware cluster management. In ASPLOS.Google Scholar
Digital Library
- Q. Deng, D. Meisner, A. Bhattacharjee, T. F. Wenisch, and R. Bianchini. 2012. CoScale: Coordinating CPU and memory system DVFS in server systems. In MICRO.Google Scholar
- Dong Hwa Kim. 2002. Parameter tuning of fuzzy neural networks by immune algorithm. In 2002 IEEE World Congress on Computational Intelligence. 2002 IEEE International Conference on Fuzzy Systems. FUZZ-IEEE’02. Proceedings (Cat. No.02CH37291).Google Scholar
- Bryan Donyanavard, Tiago Mück, Santanu Sarma, and Nikil Dutt. 2016. SPARTA: Runtime task allocation for energy efficient heterogeneous many-cores. In CODES, 2016.Google Scholar
Digital Library
- Christophe Dubach, Timothy M. Jones, and Edwin V. Bonilla. 2013. Dynamic microarchitectural adaptation using machine learning. In TACO, 2013.Google Scholar
- Christophe Dubach, Timothy M. Jones, Edwin V. Bonilla, and Michael F. P. O’Boyle. 2010. A predictive model for dynamic microarchitectural adaptivity control. In MICRO, 2010.Google Scholar
- Luc Pronzato Eric Walter. 1997. Identification of Parametric Models from Experimental Results. Springer.Google Scholar
- Hoffmann et al. 2013. A generalized software framework for accurate and efficient management of performance goals. In EMSOFT.Google Scholar
- Heirman et al. 2014. Undersubscribed threading on clustered cache architectures. In HPCA.Google Scholar
- Sui et al. 2016. Proactive control of approximate programs. In ASPLOS.Google Scholar
- Xing Fu, Khairul Kabir, and Xiaorui Wang. 2011. Cache-aware utilization control for energy efficiency in multi-core real-time systems. In ECRTS, 2011.Google Scholar
Digital Library
- Ujjwal Gupta, Raid Ayoub, Michael Kishinevsky, David Kadjo, Niranjan Soundararajan, Ugurkan Tursun, and Umit Ogras. 2017. Dynamic power budgeting for mobile systems running graphics workloads. In TMSCS, 2017.Google Scholar
- Ujjwal Gupta, Manoj Babu, Raid Ayoub, Michael Kishinevsky, Francesco Paterna, and Umit Y. Ogras. 2018. STAFF: Online learning with stabilized adaptive forgetting factor and feature selection algorithm. In 2018 55th ACM/ESDA/IEEE Design Automation Conference (DAC). IEEE, 1--6.Google Scholar
- Ujjwal Gupta, Joseph Campbell, Umit Y. Ogras, Raid Ayoub, Michael Kishinevsky, Francesco Paterna, and Suat Gumussoy. 2016. Adaptive performance prediction for integrated GPUs. In ICCAD, 2016.Google Scholar
Digital Library
- M. H. Haghbayan, A. Miele, A. M. Rahmani, P. Liljeberg, and H. Tenhunen. 2017. Performance/reliability-aware resource management for many-cores in dark silicon era. IEEE TC (2017).Google Scholar
- Nejatollahi Hamid and Salehi Mostafa E. 2018. Reliability-aware voltage scaling of multicore processors in dark silicon era. Advances in Parallel Computing 33, Big Data and HPC: Ecosystem and Convergence (2018), 242--262. DOI:https://doi.org/10.3233/978-1-61499-882-2-242Google Scholar
- Vinay Hanumaiah, Digant Desai, Benjamin Gaudette, Carole-Jean Wu, and Sarma Vrudhula. 2014. STEAM: A smart temperature and energy aware multicore controller. In TECS, 2014.Google Scholar
Digital Library
- P. Haratian, F. Safi-Esfahani, L. Salimian, and A. Nabiollahi. 2018. An adaptive and fuzzy resource management approach in cloud computing. IEEE Transactions on Cloud Computing (2018).Google Scholar
- Henry Hoffmann et al. 2011. Dynamic knobs for responsive power-aware computing. In ASPLOS.Google Scholar
- J. Hu, W. Peng, and C. Chung. 2018. Reinforcement learning for HEVC/H.265 intra-frame rate control. In 2018 IEEE International Symposium on Circuits and Systems (ISCAS).Google Scholar
- Connor Imes, Steven Hofmeyr, and Henry Hoffmann. 2018. Energy-efficient application resource scheduling using machine learning classifiers. In Proceedings of the 47th International Conference on Parallel Processing (ICPP’18).Google Scholar
Digital Library
- Engin Ipek, Onur Mutlu, José F. Martínez, and Rich Caruana. 2008. Self-optimizing memory controllers: A reinforcement learning approach. In ISCA, 2008.Google Scholar
- J. R. Jang. 1993. ANFIS: Adaptive-network-based fuzzy inference system. IEEE Transactions on Systems, Man, and Cybernetics (1993).Google Scholar
- Biresh Kumar Joardar, Ryan Gary Kim, Janardhan Rao Doppa, Partha Pratim Pande, Diana Marculescu, and Radu Marculescu. 2018. Learning-based application-agnostic 3D NoC design for heterogeneous manycore systems. IEEE Trans. Comput. (2018).Google Scholar
- H. Jung et al. 2008. Stochastic modeling of a thermally-managed multi-core system. In DAC.Google Scholar
- A. Kanduri, M. H. Haghbayan, A. M. Rahmani, P. Liljeberg, A. Jantsch, N. Dutt, and H. Tenhunen. 2016. Approximation knob: Power capping meets energy efficiency. In ICCAD, 2016.Google Scholar
Digital Library
- A Kanduri, M. H. Haghbayan, A. M. Rahmani, M. Shafique, P. Liljeberg, and A. Jantsch. 2018. dBoost: Thermal aware performance boosting through dark silicon patterning. IEEE Trans. Comput. (2018).Google Scholar
- Umair Ali Khan and Bernhard Rinner. 2014. Online learning of timeout policies for dynamic power management. ACM Trans. Embed. Comput. Syst. 13, 4 (March 2014).Google Scholar
Digital Library
- R. G. Kim, W. Choi, Z. Chen, J. R. Doppa, P. P. Pande, D. Marculescu, and R. Marculescu. 2017. Imitation learning for dynamic VFI control in large-scale manycore systems. IEEE Transactions on Very Large Scale Integration (VLSI) Systems (2017).Google Scholar
- G. J. Klir and T. A. Folge. 1988. Fuzzy Sets, Uncertainty and Information. Prentice-Hall, Englewood Cliffs, NJ, USA.Google Scholar
- G. J. Klir and B. Yuan. 1995. Fuzzy Sets and Fuzzy Logic: Theory and Applications. Prentice-Hall, Englewood Cliffs, NJ, USA.Google Scholar
- L. Ljung. 1999. System Identification: Theory for the User. Prentice Hall PTR.Google Scholar
Digital Library
- D. Mahajan et al. 2016. Towards statistical guarantees in controlling quality tradeoffs for approximate acceleration. In ISCA.Google Scholar
- R. Marculescu, U. Y. Ogras, L. Peh, N. E. Jerger, and Y. Hoskote. 2009. Outstanding research problems in NoC design: System, microarchitecture, and circuit perspectives. IEEE TCAD (2009).Google Scholar
- MathWorks. 2017. System Identification Toolbox. Technical Report. https://www.mathworks.com/products/sysid.html.Google Scholar
- Asit K. Mishra et al. 2010. CPM in CMPs: Coordinated power management in chip-multiprocessors. In SC.Google Scholar
- T. Mück, B. Donyanavard, K. Moazzemi, A. M. Rahmani, A. Jantsch, and N. Dutt. 2018. Design methodology for responsive and rrobust MIMO control of heterogeneous multicores. IEEE TMSCS (2018).Google Scholar
- Shaheryar Najam, Jameel Ahmed, Saad Masood, and Chuadhry Mujeeb Ahmed. 2019. Run-time resource management controller for power efficiency of GP-GPU architecture. IEEE Access (2019). DOI:https://doi.org/10.1109/ACCESS.2019.2901010Google Scholar
Cross Ref
- Mina Niknafs, Ivan Ukhov, Petru Eles, and Zebo Peng. 2019. Runtime resource management with workload prediction. In Proceedings of the 56th Annual Design Automation Conference 2019 (DAC’19).Google Scholar
Digital Library
- Kevin M. Passino and Stephen Yurkovich. 1997. Fuzzy Control. Addison-Wesley.Google Scholar
- P. Petrica et al. 2013. Flicker: A dynamically adaptive architecture for power limited multicore systems. In ISCA.Google Scholar
Digital Library
- Raghavendra Pothukuchi et al. 2016. A guide to design MIMO controllers for architectures. http://iacoma.cs.uiuc.edu/iacoma-papers/mimoTR.pdf (2016).Google Scholar
- Raghavendra Pradyumna Pothukuchi et al. 2016. Using multiple input, multiple output formal control to maximize resource efficiency in architectures. In ISCA.Google Scholar
- A. Prakash, H. Amrouch, M. Shafique, T. Mitra, and J. Henkel. 2016. Improving mobile gaming performance through cooperative CPU-GPU thermal management. In 2016 53nd ACM/EDAC/IEEE Design Automation Conference (DAC).Google Scholar
- A. Prakash, H. Amrouch, M. Shafique, T. Mitra, and J. Henkel. 2016. Improving mobile gaming performance through cooperative CPU-GPU thermal management. In DAC.Google Scholar
- Juan Rada-Vilela. 2018. The FuzzyLite Libraries for Fuzzy Logic Control. https://fuzzylite.com/.Google Scholar
- Ramya Raghavendra et al. 2008. No “power” struggles: Coordinated multi-level power management for the data center. In ASPLOS.Google Scholar
- A. M. Rahmani, P. Liljeberg, A. Hemani, A. Jantsch, and H. Tenhunen. 2016. The Dark Side of Silicon. Springer.Google Scholar
- Amir M. Rahmani, Bryan Donyanavard, Tiago Mück, Kasra Moazzemi, Axel Jantsch, Onur Mutlu, and Nikil Dutt. 2018. SPECTR: Formal supervisory control and coordination for many-core systems resource management. In ASPLOS.Google Scholar
Digital Library
- A. M. Rahmani, M. H. Haghbayan, A. Miele, P. Liljeberg, A. Jantsch, and H. Tenhunen. 2017. Reliability-aware runtime power management for many-core systems in the dark silicon era. IEEE TVLSI (2017).Google Scholar
- Jia Rao, Yudi Wei, Jiayu Gong, and Cheng Zhong Xu. 2011. DynaQoS: Model-free self-tuning fuzzy control of virtualized resources for QoS provisioning. IWQoS (2011). DOI:https://doi.org/10.1109/IWQOS.2011.5931341Google Scholar
- K. Rao, J. Wang, S. Yalamanchili, Y. Wardi, and Y. Handong. 2017. Application-specific performance-aware energy optimization on Android mobile devices. In HPCA.Google Scholar
- B. K. Reddy, A. K. Singh, D. Biswas, G. V. Merrett, and B. M. Al-Hashimi. 2018. Inter-cluster thread-to-core mapping and DVFS on heterogeneous multi-cores. IEEE TMSCS (2018).Google Scholar
- Leili Salimian, Faramarz Safi Esfahani, and Mohammad-Hossein Nadimi-Shahraki. 2016. An adaptive fuzzy threshold-based approach for energy and performance efficient consolidation of virtual machines. Computing (jun 2016). DOI:https://doi.org/10.1007/s00607-015-0474-5Google Scholar
- Seema Chopra, R. Mitra, and Vijay Kumar. 2004. Identification of rules using subtractive clustering with application to fuzzy controllers. In Proceedings of 2004 International Conference on Machine Learning and Cybernetics (IEEE Cat. No.04EX826).Google Scholar
Cross Ref
- Najam Shaheryar, Yasir Qadri Muhammad, Najam Zohaib, Ahmed Jameel, and N. Qadri Nadia. 2018. A fuzzy logic based power-efficient run-time reconfigurable multicore system. Chinese Journal of Electronics (May 2018). DOI:https://doi.org/10.1049/cje.2018.02.005Google Scholar
- Sina Shahosseini, Kasra Moazzemi, Amir M. Rahmani, and Nikil Dutt. 2017. Dependability evaluation of SISO control-theoretic power managers for processor architectures. NORCAS (2017).Google Scholar
- Elham Shamsa, Anil Kanduri, Amir M Rahmani, Pasi Liljeberg, Axel Jantsch, and Nikil Dutt. 2018. Goal formulation: Abstracting dynamic objectives for efficient on-chip resource allocation. In 2018 IEEE Nordic Circuits and Systems Conference (NORCAS): NORCHIP and International Symposium of System-on-Chip (SoC). IEEE, 1--4.Google Scholar
Cross Ref
- Elham Shamsa, Anil Kanduri, Amir M Rahmani, Pasi Liljeberg, Axel Jantsch, and Nikil Dutt. 2019. Goal-driven autonomy for efficient on-chip resource management: Transforming objectives to goals. In Proc. of Conf. on Design, Automation Test in Europe (DATE). IEEE.Google Scholar
Cross Ref
- Amit Kumar Singh, Alok Prakash, Karunakar Reddy Basireddy, Geoff V. Merrett, and Bashir M. Al-Hashimi. 2017. Energy-efficient run-time mapping and thread partitioning of concurrent OpenCL applications on CPU-GPU MPSoCs. ACM Trans. Embed. Comput. Syst. (Sept. 2017).Google Scholar
Digital Library
- A. K. Singh, M. Shafique, A. Kumar, and J. Henkel. 2013. Mapping on multi/many-core systems: Survey of current and emerging trends. In DAC.Google Scholar
- G. Singla, G. Kaur, A. K. Unver, and U. Y. Ogras. 2015. Predictive dynamic thermal and power management for heterogeneous mobile platforms. In 2015 Design, Automation Test in Europe Conference Exhibition (DATE).Google Scholar
- S. Skogestad and I. Postlethwaite. 2005. Multivariable Feedback Control: Analysis and Design. John Wiley 8 Sons.Google Scholar
- Janusz T. Starczewski. 2013. Defuzzification of Uncertain Fuzzy Sets. Springer Berlin Heidelberg, Berlin, Heidelberg.Google Scholar
- P Tembey et al. 2010. A case for coordinated resource management in heterogeneous multicore platforms. In ISCA.Google Scholar
- R. Teodorescu and J. Torrellas. 2008. Variation-aware application scheduling and power management for chip multiprocessors. In ISCA.Google Scholar
- K. Van Craeynest others. 2012. Scheduling heterogeneous multi-cores through performance impact estimation (PIE). In ACM SIGARCH Computer Architecture News.Google Scholar
- Augusto Vega et al. 2013. Crank it up or dial it down: Coordinated multiprocessor frequency and folding control. In MICRO.Google Scholar
- P. Viola and M. Jones. 2001. Rapid object detection using a boosted cascade of simple features. In CVPR. DOI:https://doi.org/10.1109/CVPR.2001.990517Google Scholar
- X. Wan et al. 2011. Adaptive power control with online model estimation for chip multiprocessors. IEEE TPDS (2011).Google Scholar
- Lixi Wang, Jing Xu, Ming Zhao, and José Fortes. 2011. Adaptive virtual resource management with fuzzy model predictive control. ICAC (2011). DOI:https://doi.org/10.1145/1998582.1998623Google Scholar
- Y. Wang, K. Ma, and X. Wang. 2009. Temperature-constrained power control for chip multiprocessors with online model estimation. In ISCA.Google Scholar
- Q Wu et al. 2005. Formal control techniques for power-performance management. IEEE Micro (2005).Google Scholar
- K. Yan, X. Zhang, J. Tan, and X. Fu. 2016. Redefining QoS and customizing the power management policy to satisfy individual mobile users. In MICRO.Google Scholar
- Z. Yang, L. Li, and B. Liu. 2014. Auto-tuning method of fuzzy PID controller parameter based on self-learning system. In 2014 11th International Conference on Fuzzy Systems and Knowledge Discovery (FSKD).Google Scholar
- Saehanseul Yi, Illo Yoon, Chanyoung Oh, and Youngmin Yi. 2014. Real-time integrated face detection and recognition on embedded GPGPUs. In ESTIMedia.Google Scholar
- H. Zhang et al. 2016. Maximizing performance under a power cap: A comparison of hardware, software, and hybrid techniques. In ASPLOS.Google Scholar
- Y. Zhang, J. Yao, and H. Guan. 2017. Intelligent cloud resource management with deep reinforcement learning. IEEE Cloud Computing (2017).Google Scholar
- H. J. Zimmerman. 1991. Fuzzy Set Theory--and Its Applications. Kluwer Academic Press, Boston, MA, USA.Google Scholar
Index Terms
HESSLE-FREE: Heterogeneous Systems Leveraging Fuzzy Control for Runtime Resource Management
Recommendations
SPECTR: Formal Supervisory Control and Coordination for Many-core Systems Resource Management
ASPLOS '18: Proceedings of the Twenty-Third International Conference on Architectural Support for Programming Languages and Operating SystemsResource management strategies for many-core systems need to enable sharing of resources such as power, processing cores, and memory bandwidth while coordinating the priority and significance of system- and application-level objectives at runtime in a ...
SPECTR: Formal Supervisory Control and Coordination for Many-core Systems Resource Management
ASPLOS '18Resource management strategies for many-core systems need to enable sharing of resources such as power, processing cores, and memory bandwidth while coordinating the priority and significance of system- and application-level objectives at runtime in a ...
A timeenergy performance analysis of MapReduce on heterogeneous systems with GPUs
Motivated by the explosion of Big Data analytics, performance improvements in low-power (wimpy) systems and the increasing energy efficiency of GPUs, this paper presents a timeenergy performance analysis of MapReduce on heterogeneous systems with GPUs. ...






Comments