Scaling Poisson Solvers on Many Cores via MMEwald. Wu, M., Wu, Y., Shang, H., Liu, Y., Cui, H., Li, F., Duan, X., Zhang, Y., & Feng, X. IEEE Transactions on Parallel and Distributed Systems, 33(8):1888–1901, aug, 2022. ![Scaling Poisson Solvers on Many Cores via MMEwald [link]](https://bibbase.org/img/filetypes/link.svg "link")
Paper doi abstract bibtex The Poisson solver for the calculation of the electrostatic potential is an essential primitive in quantum mechanics calculations. In this article, we adopt the Ewald method and propose a highly-optimized and scalable framework for Poisson solver, MMEwald, on the new generation Sunway supercomputer, capable of utilizing the collection of 390-core accelerators it uses. The MMEwald is based on a grid adapted cut-plane approach to partition the points into batches and distribute the batch to the processors. Furthermore, we propose a set of architecture-specific optimizations to efficiently utilize the memory bandwidth and computation capacity of the supercomputer. Experimental results demonstrate the efficiency of the MMEwald in providing strong and weak scaling performance.
@article{Wu2021,
abstract = {The Poisson solver for the calculation of the electrostatic potential is an essential primitive in quantum mechanics calculations. In this article, we adopt the Ewald method and propose a highly-optimized and scalable framework for Poisson solver, MMEwald, on the new generation Sunway supercomputer, capable of utilizing the collection of 390-core accelerators it uses. The MMEwald is based on a grid adapted cut-plane approach to partition the points into batches and distribute the batch to the processors. Furthermore, we propose a set of architecture-specific optimizations to efficiently utilize the memory bandwidth and computation capacity of the supercomputer. Experimental results demonstrate the efficiency of the MMEwald in providing strong and weak scaling performance.},
author = {Wu, Mingchuan and Wu, Yangjun and Shang, Honghui and Liu, Ying and Cui, Huimin and Li, Fang and Duan, Xiaohui and Zhang, Yunquan and Feng, Xiaobing},
doi = {10.1109/TPDS.2021.3127138},
file = {:C\:/Users/Administrator/AppData/Local/Mendeley Ltd./Mendeley Desktop/Downloaded/Wu et al. - 2022 - Scaling Poisson Solvers on Many Cores via MMEwald.pdf:pdf},
issn = {15582183},
journal = {IEEE Transactions on Parallel and Distributed Systems},
keywords = {Architecture-specific optimizations,Many-core processor,Poisson solver},
month = {aug},
number = {8},
pages = {1888--1901},
title = {{Scaling Poisson Solvers on Many Cores via MMEwald}},
url = {https://ieeexplore.ieee.org/document/9611019/},
volume = {33},
year = {2022}
}
Downloads: 0
{"_id":"EmGJZAunFrgGZHCPy","bibbaseid":"wu-wu-shang-liu-cui-li-duan-zhang-etal-scalingpoissonsolversonmanycoresviammewald-2022","author_short":["Wu, M.","Wu, Y.","Shang, H.","Liu, Y.","Cui, H.","Li, F.","Duan, X.","Zhang, Y.","Feng, X."],"bibdata":{"bibtype":"article","type":"article","abstract":"The Poisson solver for the calculation of the electrostatic potential is an essential primitive in quantum mechanics calculations. In this article, we adopt the Ewald method and propose a highly-optimized and scalable framework for Poisson solver, MMEwald, on the new generation Sunway supercomputer, capable of utilizing the collection of 390-core accelerators it uses. The MMEwald is based on a grid adapted cut-plane approach to partition the points into batches and distribute the batch to the processors. Furthermore, we propose a set of architecture-specific optimizations to efficiently utilize the memory bandwidth and computation capacity of the supercomputer. Experimental results demonstrate the efficiency of the MMEwald in providing strong and weak scaling performance.","author":[{"propositions":[],"lastnames":["Wu"],"firstnames":["Mingchuan"],"suffixes":[]},{"propositions":[],"lastnames":["Wu"],"firstnames":["Yangjun"],"suffixes":[]},{"propositions":[],"lastnames":["Shang"],"firstnames":["Honghui"],"suffixes":[]},{"propositions":[],"lastnames":["Liu"],"firstnames":["Ying"],"suffixes":[]},{"propositions":[],"lastnames":["Cui"],"firstnames":["Huimin"],"suffixes":[]},{"propositions":[],"lastnames":["Li"],"firstnames":["Fang"],"suffixes":[]},{"propositions":[],"lastnames":["Duan"],"firstnames":["Xiaohui"],"suffixes":[]},{"propositions":[],"lastnames":["Zhang"],"firstnames":["Yunquan"],"suffixes":[]},{"propositions":[],"lastnames":["Feng"],"firstnames":["Xiaobing"],"suffixes":[]}],"doi":"10.1109/TPDS.2021.3127138","file":":C\\:/Users/Administrator/AppData/Local/Mendeley Ltd./Mendeley Desktop/Downloaded/Wu et al. - 2022 - Scaling Poisson Solvers on Many Cores via MMEwald.pdf:pdf","issn":"15582183","journal":"IEEE Transactions on Parallel and Distributed Systems","keywords":"Architecture-specific optimizations,Many-core processor,Poisson solver","month":"aug","number":"8","pages":"1888–1901","title":"Scaling Poisson Solvers on Many Cores via MMEwald","url":"https://ieeexplore.ieee.org/document/9611019/","volume":"33","year":"2022","bibtex":"@article{Wu2021,\nabstract = {The Poisson solver for the calculation of the electrostatic potential is an essential primitive in quantum mechanics calculations. In this article, we adopt the Ewald method and propose a highly-optimized and scalable framework for Poisson solver, MMEwald, on the new generation Sunway supercomputer, capable of utilizing the collection of 390-core accelerators it uses. The MMEwald is based on a grid adapted cut-plane approach to partition the points into batches and distribute the batch to the processors. Furthermore, we propose a set of architecture-specific optimizations to efficiently utilize the memory bandwidth and computation capacity of the supercomputer. Experimental results demonstrate the efficiency of the MMEwald in providing strong and weak scaling performance.},\nauthor = {Wu, Mingchuan and Wu, Yangjun and Shang, Honghui and Liu, Ying and Cui, Huimin and Li, Fang and Duan, Xiaohui and Zhang, Yunquan and Feng, Xiaobing},\ndoi = {10.1109/TPDS.2021.3127138},\nfile = {:C\\:/Users/Administrator/AppData/Local/Mendeley Ltd./Mendeley Desktop/Downloaded/Wu et al. - 2022 - Scaling Poisson Solvers on Many Cores via MMEwald.pdf:pdf},\nissn = {15582183},\njournal = {IEEE Transactions on Parallel and Distributed Systems},\nkeywords = {Architecture-specific optimizations,Many-core processor,Poisson solver},\nmonth = {aug},\nnumber = {8},\npages = {1888--1901},\ntitle = {{Scaling Poisson Solvers on Many Cores via MMEwald}},\nurl = {https://ieeexplore.ieee.org/document/9611019/},\nvolume = {33},\nyear = {2022}\n}\n\n\n\n","author_short":["Wu, M.","Wu, Y.","Shang, H.","Liu, Y.","Cui, H.","Li, F.","Duan, X.","Zhang, Y.","Feng, X."],"key":"Wu2021","id":"Wu2021","bibbaseid":"wu-wu-shang-liu-cui-li-duan-zhang-etal-scalingpoissonsolversonmanycoresviammewald-2022","role":"author","urls":{"Paper":"https://ieeexplore.ieee.org/document/9611019/"},"keyword":["Architecture-specific optimizations","Many-core processor","Poisson solver"],"metadata":{"authorlinks":{}}},"bibtype":"article","biburl":"https://bibbase.org/f/rxmgC5ZwPjrBPCwT8/publication_2010-2022.bib","dataSources":["uRC7qTgdgjYp52rwx","2252seNhipfTmjEBQ","ip3orsyTWaDmn4rSF","EKd5Jwkp9dGrKqYfk"],"keywords":["architecture-specific optimizations","many-core processor","poisson solver"],"search_terms":["scaling","poisson","solvers","many","cores","via","mmewald","wu","wu","shang","liu","cui","li","duan","zhang","feng"],"title":"Scaling Poisson Solvers on Many Cores via MMEwald","year":2022}