Linear systems solvers for distributed-memory machines with GPU accelerators

DOI10.1007/978-3-030-29400-7_35MaRDI QIDQ3297578zbMATH OpenOpenAlexFDO

Authors Jakub Kurzak, Mark Ralph Gates, Ali Charara, Asim Yarkhan, Ichitaro Yamazaki, Jack Dongarra

Publication date 20 July 2020

Published in Lecture Notes in Computer Science (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1007/978-3-030-29400-7_35

zbMATH Keywords

Cholesky factorizationGPU accelerationlinear algebraLU factorizationdistributed memorylinear systems of equations

Mathematics Subject Classification ID

Direct numerical methods for linear systems and matrix inversion (65F05) Parallel numerical computation (65Y05) Software, source code, etc. for problems pertaining to numerical analysis (65-04) Numerical algorithms for specific classes of architectures (65Y10)

Recommendations

• Accelerating preconditioned iterative linear solvers on GPU
• Performance models and workload distribution algorithms for optimizing a hybrid CPU-GPU multifrontal solver
• Effective minimally-invasive GPU acceleration of distributed sparse matrix factorization
• Exposing fine-grained parallelism in algebraic multigrid methods
• A data-parallel ILUPACK for sparse general and symmetric indefinite linear systems

Cites work

• scientific article; zbMATH DE number 2089172 (Why is no real title available?)
• A high performance QDWH-SVD solver using hardware accelerators
• A recursive formulation of Cholesky factorization of a matrix in packed storage
• Parallel and cache-efficient in-place matrix storage format conversion
• ScaLAPACK Users' Guide
• Scaling LAPACK panel operations using parallel cache assignment

Cited in

(6)

• A condensation-based application of Cramer's rule for solving large-scale linear systems
• Considerations on the Implementation and Use of Anderson Acceleration on Distributed Memory and GPU-based Parallel Computers
• Performance models and workload distribution algorithms for optimizing a hybrid CPU-GPU multifrontal solver
• A flexible CUDA LU-based solver for small, batched linear systems
• Towards dense linear algebra for hybrid GPU accelerated manycore systems
• An error correction solver for linear systems: evaluation of mixed precision implementations

Describes a project that uses

Uses Software

• Slate
• ScaLAPACK
• LAPACK

This page was built for publication: Linear systems solvers for distributed-memory machines with GPU accelerators

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3297578)