PPT-Addressing the straggler problem for iterative convergent parallel ML

Aaron Harlap Henggang Cui Wei Dai Jinliang Wei Gregory R Ganger Phillip B Gibbons Garth A Gibson Eric P Xing One slide overview Workers are a single thread on a machine Stragglers are bad. Computations. K-means. Performance of K-Means. Smith Waterman is a non iterative case and of course runs fine. Matrix Multiplication . 64 cores. Square blocks Twister. Row/Col . decomp. Twister. Efficient and scalable architectures to perform pleasingly parallel, MapReduce and iterative data intensive computations on cloud environments. Thilina. . Gunarathne. (tgunarat@indiana.edu). Advisor : . Richard . Peng. Joint with Michael Cohen (MIT), . Rasmus. . Kyng. (Yale), . Jakub. . Pachocki. (CMU), and . Anup. . Rao. (Yale). MIT. CMU theory seminar, April 5, 2014. Random Sampling. Collection of many objects. With Team 3. Divergent Thinking. Divergent thinking.  is a thought process or method used to generate creative ideas by exploring many possible . solutions. . Divergent . thinking typically occurs in a spontaneous, free-flowing, 'non-linear' manner, such that many ideas are generated in an emergent cognitive fashion. . Efficient and scalable architectures to perform pleasingly parallel, MapReduce and iterative data intensive computations on cloud environments. Thilina. . Gunarathne. (tgunarat@indiana.edu). Advisor : . Tsung. Yu. , Sheng-Han . Yeh. , and . Tsung. -Yi Ho. jidung@eda.csie.ncku.edu.tw. http://eda.csie.ncku.edu.tw. Electronic Design Automation Laboratory. Department of Computer Science and Information Engineering. Iterative AS and PA in Energy Efficient MIMO. IEEE ICC 2014 . 1. Is there a promising way? . A. n. I. terative. A. lgorithm. . for. J. oint. A. ntenna. S. election. . and. P. ower. A. daptation. Dense A:. Gaussian elimination with partial pivoting (LU). Same flavor as matrix * matrix, but more complicated. Sparse A:. Gaussian elimination – Cholesky, LU, etc.. Graph algorithms. Sparse A:. Department of Computer Science and Information Engineering. National Cheng Kung University, Tainan Taiwan. Tsung-Wei Huang, Tsung-Yi Ho, and Krishnendu Chakrabarty. Department of Electrical and Computer Engineering. Computations. K-means. Performance of K-Means. Smith Waterman is a non iterative case and of course runs fine. Matrix Multiplication . 64 cores. Square blocks Twister. Row/Col . decomp. Twister. Iterative Local Searches. Martin . Burtscher. 1. and Hassan Rabeti. 2. 1. Department of Computer Science, Texas State University-San Marcos. 2. Department of Mathematics, Texas State University-San Marcos. Iterative Local Searches. Martin . Burtscher. 1. and Hassan Rabeti. 2. 1. Department of Computer Science, Texas State University-San Marcos. 2. Department of Mathematics, Texas State University-San Marcos. Table of Contents. Overview of Evolutionary Computation. Programming Strategies. Related Architecture. Efficiency Comparison. Implementations. Conclusion. Origins. Based on the study of evolution by Charles Darwin. We will Study …. Concurrency . Parallelism. Distributed computing. Evaluation. Assignments 40%, . Minor-1 15%, . Minor-2 15%, . Major 30%. Plagiarism is unacceptable. Offenders will be . penalized by a failing.