Search academic texts

Information × Registration Number 0219U003121, 0117U006182 , R & D reports Title To develop parallel algorithms for solving problems of linear algebra with sparse data structures on computers with Intel Xeon Phi processors popup.stage_title Head Sidoruk Volodymyr, Registration Date 22-01-2019 Organization V.M.Glushkov Institute of Cybernetics, National Academy of Science of Ukraine popup.description2 Methods and computer algorithms for reordering and decomposing sparse matrices, layout and storage schemes for sparse data on computers with parallel computing organization and Intel Xeon Phi processors x200 series have been developed. Methods and computer algorithms of factorization of sparse matrices on parallel computer systems are developed. Iterative methods and computer algorithms of parallel computations for research and solving systems of linear algebraic equations with sparse matrices are developed. The methods and computer algorithms of parallel computing for the study and solving algebraic problem of eigenvalues of sparse matrices are developed. The experimental research of the offered methods and algorithms on solving a series of test and demonstration tasks on parallel computer systems with processors Intel Xeon Phi series x200 is conducted. Together with specialists from the Institute of Mechanics named after SP Tymoshenko National Academy of Sciences of Ukraine on the basis of parallel computations the problem of stability of composite material is solved, considerable acceleration of calculations is obtained. Product Description popup.authors Сидорук В.А. Чистяков О.В. popup.nrat_date 2020-04-02 Close