隨著漸增的多樣性需求，異質多核心系統已經變成一種流行的架構。其中一個具有象徵意義的系統是big.LITTLE技術，它結合了高效能的大核心以及節能的小核心，目的是同時實現高效能與低耗能的需求。對於這樣一個系統而言，雖然人們相信最節能的執行方式是把程式安排給小核心，然而沒有考慮大核心的閒置功耗會擊垮這個簡單的策略。
在此篇論文中我們提出一個有效的排程演算法，針對big.LITTLE異質多核心系統，最小化系統的能量消耗。首先我們建立一個整數線性規劃的模型使得系統的能量消耗最小，並且用Matlab來解這個模型。我們從最佳解中觀察它們的特性，並依此特性提出了一個建立在Longest Time First啟發式演算法與區域搜尋法為基礎上的演算法。
我們在ODROID-XU3開發板上實作並測試我們的模型與演算法，此開發板的處理器是由4個小的in-order Cortex-A7核心與4個大的out-of-order Cortex-A15核心。針對各種不同的workload，實驗結果顯示我們提出的排程演算法跟最佳解平均差不到1%。而且我們的方法會比直接解整數線性規劃模型更有效率的多。

With the increasing diversity of demands, heterogeneous multi-core systems have become a popular architecture. One of the representative systems is the big.LITTLE technology, which combines high-performance big cores with the energy efficient little cores to achieve both high performance and low power. For such system, although people believe the most energy efficient way to execute tasks is scheduling them on little cores, the overlooked idle power consumption of big cores could overwhelm this simple energy saving strategy. In this thesis, we propose an efficient scheduling algorithm to minimize the energy consumption of the big.LITTLE heterogeneous multi-core systems. We first build an integer linear programming model that minimize the total energy, and solve it using Matlab. From the optimal solutions we observe their properties, base on which, we derive an algorithm using the longest time first heuristic and the local search approach. We implemented and tested our model and algorithm on an ODROID-XU3 development board, which consisting of four small in-order Cortex-A7 and four big out-of-order Cortex-A15 cores. The experimental results show that for the total energy consumption, the differences between the results obtained by our scheduling algorithm and the optimal solution are within 1\% on the average for various workloads. But our algorithm is much more efficient than solving the integer linear programming model.

Chinese Abstract i
Abstract ii
Contents iv
List of Figures vi
List of Tables vii
1 Introduction 1
2 Energy Minimization Model 4
2.1 An Illustrative Example . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.2 Integer Linear Programming Model . . . . . . . . . . . . . . . . . . . 7
3 Energy Minimization Scheduling Algorithm 12
3.1 Core-Dominant and Idle-Dominant . . . . . . . . . . . . . . . . . . . 12
3.2 Idle-Dominant Factor First Algorithm . . . . . . . . . . . . . . . . . . 13
3.2.1 Optimizing IDF . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.2.2 Optimizing CDF . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.3 Compute-Dominant Factor First Algorithm . . . . . . . . . . . . . . . 18
3.3.1 Optimizing CDF . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.3.2 Optimizing IDF . . . . . . . . . . . . . . . . . . . . . . . . . . 19
4 Experiments and Results 21
4.1 Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.2 Variations in Idle Power . . . . . . . . . . . . . . . . . . . . . . . . . 22
4.3 Model and EMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
4.4 Measured and Computed Results of EMS . . . . . . . . . . . . . . . . 27
4.5 Comparison with Other Scheduling Policies . . . . . . . . . . . . . . . 28
5 Conclusion 33
v

[1] Michela Becchi and Patrick Crowley. Dynamic thread assignment on heterogeneous multiprocessor architectures. In Proceedings of the 3rd Conference on
Computing Frontiers, CF '06, pages 29-40, New York, NY, USA, 2006. ACM.
[2] Trevor E. Carlson, Wim Heirman, and Lieven Eeckhout. Sniper: Exploring the level of abstraction for scalable and accurate parallel multi-core simulation. In
Proceedings of 2011 International Conference for High Performance Computing, Networking, Storage and Analysis, SC '11, pages 52:1{52:12, New York, NY, USA, 2011. ACM.
[3] Jian Chen and Lizy K. John. Ecient program scheduling for heterogeneous multi-core processors. In Proceedings of the 46th Annual Design Automation Conference, DAC '09, pages 927-930, New York, NY, USA, 2009. ACM.
[4] N. Chitlur, G. Srinivasa, S. Hahn, P. K. Gupta, D. Reddy, D. Koufaty, P. Brett, A. Prabhakaran, L. Zhao, N. Ijih, S. Subhaschandra, S. Grover, X. Jiang, and R. Iyer. Quickia: Exploring heterogeneous architectures on real prototypes. In IEEE International Symposium on High-Performance Comp Architecture, pages 1-8, Feb 2012.
[5] Jason Cong and Bo Yuan. Energy-ecient scheduling on heterogeneous multi-core architectures. In Proceedings of the 2012 ACM/IEEE International Symposium on Low Power Electronics and Design, ISLPED '12, pages 345-350, New York, NY, USA, 2012. ACM.
[6] Rakesh Kumar, Keith I. Farkas, Norman P. Jouppi, Parthasarathy Ranganathan, and Dean M. Tullsen. Single-isa heterogeneous multi-core architectures: The potential for processor power reduction. In Proceedings of the 36th Annual IEEE/ACM International Symposium on Microarchitecture, MICRO 36, pages 81-, Washington, DC, USA, 2003. IEEE Computer Society.
[7] Nagesh B. Lakshminarayana, Jaekyu Lee, and Hyesoon Kim. Age based scheduling for asymmetric multiprocessors. In Proceedings of the Conference on High Performance Computing Networking, Storage and Analysis, SC '09, pages 25:1-25:12, New York, NY, USA, 2009. ACM.
[8] G. Liu, J. Park, and D. Marculescu. Dynamic thread mapping for high-performance, power-ecient heterogeneous many-core systems. In 2013 IEEE 31st International Conference on Computer Design (ICCD), pages 54-61, Oct 2013.
[9] Vinicius Petrucci, Orlando Loques, and Daniel Mosse. Lucky scheduling for
energy-ecient heterogeneous multi-core systems. In HotPower, 2012.
[10] Juan Carlos Saez, Daniel Shelepov, Alexandra Fedorova, and Manuel Prieto. Leveraging workload diversity through os scheduling to maximize performance on single-isa heterogeneous multicore systems. J. Parallel Distrib. Comput., 71(1):114-131, January 2011.
[11] L. Sawalha and R. D. Barnes. Energy-ecient phase-aware scheduling for heterogeneous multicore processors. In 2012 IEEE Green Technologies Conference, pages 1-6, April 2012.
[12] Daniel Shelepov, Juan Carlos Saez Alcaide, Stacey Jeery, Alexandra Fedorova, Nestor Perez, Zhi Feng Huang, Sergey Blagodurov, and Viren Kumar. Hass: A scheduler for heterogeneous multicore systems. SIGOPS Oper. Syst. Rev., 43(2):66-75, April 2009.
[13] Kenzo Van Craeynest, Aamer Jaleel, Lieven Eeckhout, Paolo Narvaez, and Joel Emer. Scheduling heterogeneous multi-cores through performance impact estimation (pie). In Proceedings of the 39th Annual International Symposium on Computer Architecture, ISCA '12, pages 213{224, Washington, DC, USA, 2012. IEEE Computer Society.
[14] ODROID-XU3 http://www.hardkernel.com/main/products/prdt_info.php?g_code=g140448267127.
[15] ODROID-U2 http://www.hardkernel.com/main/products/prdt_info.php?g_code=G135341370451.
[16] TI INA-231 http://www.hardkernel.com/main/products/prdt_info.php?g_code=G135341370451.
[17] Nvidia Tegra X1 http://www.nvidia.com/object/tegra-x1-processor.html.
[18] MediaTek Helio X25 http://mediatek-helio.com/x20/.
[19] Samsung Exynos 8890 http://www.samsung.com/semiconductor/minisite/Exynos/w/solution/mod_ap/8890/.
[20] Intel Skylake http://ark.intel.com/products/codename/37572/Skylake#@All.
[21] SPEC CPU2006 https://www.spec.org/cpu2006/.
[22] ARM Streamline Performance Analyzer https://developer.arm.com/products/software-development-tools/ds-5-development-studio/streamline.