2016 Dense Linear Algebra Software Packages Survey

Title2016 Dense Linear Algebra Software Packages Survey
Publication TypeTech Report
Year of Publication2016
AuthorsDongarra, J., J. Demmel, J. Langou, and J. Langou
Technical Report Series TitleUniversity of Tennessee Computer Science Technical Report
NumberUT-EECS-16-744 / LAWN 290
Date Published2016-09
InstitutionUniversity of Tennessee
Abstract

The 2016 Dense Linear Algebra Software Packages Survey was administered from January 1st 2016 to April 12 2016. 234 respondents answered the survey. The survey was advertised directly to the Linear Algebra community via our LAPACK/ScaLAPACK forum, NA Digest and we also directly contacted vendors and linear algebra experts. The breakdown of respondents was: 74% researchers or scientists, 25% were Principal Investigators and 25% Software maintainers or System administrators. The goal of the survey was to get the Linear Algebra community opinion and provide input on dense linear algebra software packages, in particular LAPACK, ScaLAPACK, PLASMA and MAGMA. The ultimate purpose of the survey was to improve these libraries to benefit our user community. The survey would allow the team to prioritize the many possible improvements that could be done. We also asked input from users accessing these libraries via 3rd party interfaces, for example MATLAB, Intel’s MKL, Python’s NumPy, AMD's ACML, and many others.

Project Tags: 
External Publication Flag: