Lattice QCD with domain decomposition on Intel® Xeon Phi™ co-processors

• Published in: Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis pp. 69 - 80
• Main Authors: Heybrock, Simon, Joó, Bálint, Kalamkar, Dhiraj D., Smelyanskiy, Mikhail, Vaidyanathan, Karthikeyan, Wettig, Tilo, Dubey, Pradeep
• Format: Conference Proceeding
• Language: English
• Published: Piscataway, NJ, USA IEEE Press 16.11.2014; IEEE
• Series: ACM Conferences
• Subjects: and very la; Applied computing > Physical sciences and engineering > Physics; Computing methodologies > Symbolic and algebraic manipulation > Symbolic and algebraic algorithms > Linear algebra algorithms; Domain decomposition; G.1.3 [Numerical Analysis]: Numerical Linear Algebra Sparse; Gold; Intel® Xeon Phi coprocessor; Jacobian matrices; Lattice QCD Categories and subject descriptors: D.3.4 [Programming Languages]: Processors Optimization; Lattices; Layout; Linear systems; Mathematics of computing > Mathematical analysis > Numerical analysis > Computations on matrices; Mathematics of computing > Mathematical software; Prefetching; Software and its engineering > Software notations and tools > Compilers; structured; Vectors; Xeon Phi; domain decomposition; Intel; coprocessor; lattice QCD
• ISBN: 1479955000; 9781479955008
• ISSN: 2167-4329
• DOI: 10.1109/SC.2014.11

The gap between the cost of moving data and the cost of computing continues to grow, making it ever harder to design iterative solvers on extreme-scale architectures. This problem can be alleviated by alternative algorithms that reduce the amount of data movement. We investigate this in the context of Lattice Quantum Chromodynamics and implement such an alternative solver algorithm, based on domain decomposition, on Intel® Xeon Phi™ co-processor (KNC) clusters. We demonstrate close-to-linear on-chip scaling to all 60 cores of the KNC. With a mix of single- and half-precision the domain-decomposition method sustains 400-500 Gflop/s per chip. Compared to an optimized KNC implementation of a standard solver [1], our full multi-node domain-decomposition solver strong-scales to more nodes and reduces the time-to-solution by a factor of 5.