MPSoC Communication Optimization Strategy Based on Integer Linear Programming

Abstract

Abstract:

Abstract： To solve the problem of the overuse of communication optimization techniques—communication pipeline and message aggregation—in MPSoC (Multiprocessor System-on-chip), which may increase the synchronization cost and lead to system performance degradation, an ILP (integer linear programming) based strategy was proposed. The strategy can employ communication pipeline and message aggregation techniques appropriately, considering communication optimization as well as synchronization optimization, to achieve optimal system performance. Experimental results show that the proposed strategy can take the advantage of the communication optimization techniques and obtain significant performance improvements.

Key words: communication optimization, integer linear programming (ILP), Simulink

HUANG Kai1a,ZHANG Xiaoxu1b,ZHANG Xiaomeng1b,XIU Siwen2, MEI Lingqi1a,YU Min1a,YAN Xiaolang1b. MPSoC Communication Optimization Strategy Based on Integer Linear Programming[J]. Journal of Shanghai Jiaotong University, 2015, 49(02): 184-190.

References

［1］Popovici K, Rousseau F, Jerraya A A, et al. Embedded software design and programming of multiprocessor systemonchip: Simulink and system C case studies ［M］.New York: Springer, 2010: 116.

［2］Huang K, Haid W, Bcacivarov I, et al. Embedding formal performance analysis into the design cycle of MPSoCs for realtime streaming applications ［J］. ACM Transactions on Embedded Computing Systems (TECS), 2012, 11(1): 8.

［3］Castrillon J, Leupers R, Ascheid G. Maps: Mapping concurrent dataflow applications to heterogeneous MPSoCs ［J］. Industrial Informatics, IEEE Transactions on, 2013, 9(1): 527545.

［4］Castrillon J, Tretter A, Leupers R, et al. Communicationaware mapping of KPN applications onto heterogeneous MPSoCs ［C］∥Proceedings of the 49th Annual Design Automation Conference. New York, USA: ACM, 2012: 12661271.

［5］Kahn G, Macqueen D. Coroutines and networks of parallel processes ［R］. France: HAL,1976.

［6］The Mathworks, Inc. Simulink ［EB/OL］. (19941023)［20140706］. http:∥cn.mathworks.com/products/simulink/.

［7］Grepl R. Realtime control prototyping in MATLAB/Simulink: Review of tools for research and education in mechatronics ［C］∥Mechatronics (ICM), 2011 IEEE International Conference on. New York, USA: IEEE, 2011: 881886.

［8］Cha M, Kim K H, Lee C J, et al. Deriving highperformance realtime multicore systems based on Simulink applications ［C］∥Dependable, Autonomic and Secure Computing (DASC)， 2011 IEEE Ninth International Conference on. New York, USA: IEEE, 2011: 267274.

［9］Canedo A, Yoshizawa T, Komatsu H. Automatic parallelization of simulink applications ［C］∥Proceedings of the 8th Annual IEEE/ACM International Symposium on Code Generation and Optimization. New York, USA: ACM, 2010: 151159.

［10］Canedo A, Yoshizawa T, Komatsu H. Skewed pipelining for parallel simulink simulations ［C］∥Design, Automation & Test in Europe Conference & Exhibition (DATE). New York, USA: IEEE, 2010: 891896.

［11］Han S I, Chae S I, Brisolara L, et al. Simulinkbased heterogeneous multiprocessor SoC design flow for mixed hardware/software refinement and simulation ［J］. Integration, the VLSI Journal, 2009, 42(2): 227245.

［12］Brisolara L, Han S I, Guerin X, et al. Reducing finegrain communication overhead in multithread code generation for heterogeneous MPSoC ［C］∥Proceedings of the 10th International Workshop on Software & Compilers for Embedded Systems. New York, USA: ACM, 2007: 8189.

［13］Yan R, Huang K, Yu M, et al. Communication pipelining for code generation from simulink models ［C］∥Trust, Security and Privacy in Computing and Communications (TrustCom), 2013 12th IEEE International Conference on. New York, USA: IEEE, 2013: 18931900.

［14］Chen D S, Batson R G, Dang Y. Applied integer programming: Modeling and solution ［M］. New Jersey: John Wiley & Sons, 2011:38.

［15］Han S I, Baghdadi A, Bonaciu M, et al. An efficient scalable and flexible data transfer architecture for multiprocessor SoC with massive distributed memory ［C］∥Proceedings of the 41st Annual Design Automation Conference. New York, USA: ACM, 2004: 250255.

［16］CSKY Microsystems. CSKY嵌入式CPU之CK803 ［EB/OL］.(20120912)［20140723］.http:∥www.csky.com/product.php?typeid=84.

[1]	WANG Ning, FU Yunpeng, LI Ting, LI Tie, YI Ping. Optimization of Control Scheme for Large Flow Seawater Cooling System Based on FloMaster-Simulink Co-Simulation [J]. Journal of Shanghai Jiao Tong University, 2022, 56(3): 379-385.
[2]	QI Yuanqu，YU Yuefeng. Numerical Simulation and Analysis of Leakage Process in Twin Screw Expander [J]. Journal of Shanghai Jiaotong University, 2016, 50(04): 496-501.
[3]	LIN Chao1,JI Jiuxiang1,CAI Lizhong1,SHAO Jiming2. Bond Graph Method for Analysis of Motion Along Coordinate Axis of Micro/Nano Transmission Platform [J]. Journal of Shanghai Jiaotong University, 2015, 49(09): 1311-1318.