Reducing floating point error in dot product using the superblock family of algorithms

DOI10.1137/070679946MaRDI QIDQ110775zbMATHOpenAlexFDO

Authors Anthony M. Castaldo, R. Clint Whaley, Anthony T. Chronopoulos, Anthony M. Castaldo, R. Clint Whaley, Anthony T. Chronopoulos

Publication date January 2009

Published in SIAM Journal on Scientific Computing (Search for Journal in Brave)

Full work available at URL https://semanticscholar.org/paper/2483c58592a7651e8f82db540cf19e9ce219f9f2

zbMATH Keywords

algorithmserror boundsnumerical examplesATLASBLAScomputational performancedot producterror analysiserror behaviorerror-reducing algorithmsinner productmemory usagesuperblock

Mathematics Subject Classification ID

Complexity and performance of numerical algorithms (65Y20) Roundoff error (65G50)

Recommendations

• A fast dot-product algorithm with minimal rounding errors
• Accurate Sum and Dot Product
• Ultimately fast accurate summation
• Compensated summation and dot product algorithms for floating-point vectors on parallel architectures: error bounds, implementation and application in the Krylov subspace methods
• Error estimation of floating-point summation and dot product

Cited in

(9)

• Floating-point arithmetic
• A fast dot-product algorithm with minimal rounding errors
• Verified numerical computations for large-scale linear systems.
• A Class of Fast and Accurate Summation Algorithms
• A New Approach to Probabilistic Rounding Error Analysis
• PreciseSums
• Numerical stability of algorithms at extreme scale and low precisions
• Minimizing synchronizations in sparse iterative solvers for distributed supercomputers
• Sharper probabilistic backward error analysis for basic linear algebra kernels with random data

Describes a project that uses

Uses Software

• ATLAS
• ScaLAPACK

This page was built for publication: Reducing floating point error in dot product using the superblock family of algorithms

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q110775)