Dependable direct solutions for linear systems using a little extra precision

Dependable direct solutions for linear systems using a little extra precision. Riedy, E. J. CSE Seminar at Georgia Institute of Technology, August, 2009. Invited presentation

Paper abstract bibtex

Solving a square linear system $Ax=b$ often is considered a black box. It's supposed to "just work," and failures often are blamed on the original data or subtleties of floating-point. Now that we have an abundance of cheap computations, however, we can do much better. A little extra precision in just the right places produces accurate solutions cheaply or demonstrates when problems are too hard to solve without significant cost. This talk will outline the method, iterative refinement with a new twist; the benefits, small backward and forward errors; and the trade-offs and unexpected benefits.

@misc{gt09,
  file = {material/gt-2009-08-21.pdf},
  author = {E. Jason Riedy},
  title = {Dependable direct solutions for linear systems using a little extra precision},
  howpublished = {CSE Seminar at Georgia Institute of Technology},
  dom = 21,
  month = aug,
  year = 2009,
  url = {http://hdl.handle.net/1853/29795},
  opttags = {linear algebra; floating point; lapack},
  note = {Invited presentation},
  abstract = {Solving a square linear system $Ax=b$ often is considered a black box. It's supposed to "just work," and failures often are blamed on the original data or subtleties of floating-point. Now that we have an abundance of cheap computations, however, we can do much better. A little extra precision in just the right places produces accurate solutions cheaply or demonstrates when problems are too hard to solve without significant cost. This talk will outline the method, iterative refinement with a new twist; the benefits, small backward and forward errors; and the trade-offs and unexpected benefits.},
  projtag = {lapack, sparse-methods, ieee754},
  keywords = {linear algebra, sparse matrix, foating point, lapack},
  ejr-proj = {linear-algebra}
}

Downloads: 0

{"_id":"iKvKjez46svpehe4c","bibbaseid":"riedy-dependabledirectsolutionsforlinearsystemsusingalittleextraprecision-2009","downloads":0,"creationDate":"2018-08-21T18:11:53.630Z","title":"Dependable direct solutions for linear systems using a little extra precision","author_short":["Riedy, E. J."],"year":2009,"bibtype":"misc","biburl":"http://lovesgoodfood.com/jason/CV/ejr.bib","bibdata":{"bibtype":"misc","type":"misc","file":"material/gt-2009-08-21.pdf","author":[{"firstnames":["E.","Jason"],"propositions":[],"lastnames":["Riedy"],"suffixes":[]}],"title":"Dependable direct solutions for linear systems using a little extra precision","howpublished":"CSE Seminar at Georgia Institute of Technology","dom":"21","month":"August","year":"2009","url":"http://hdl.handle.net/1853/29795","opttags":"linear algebra; floating point; lapack","note":"Invited presentation","abstract":"Solving a square linear system $Ax=b$ often is considered a black box. It's supposed to \"just work,\" and failures often are blamed on the original data or subtleties of floating-point. Now that we have an abundance of cheap computations, however, we can do much better. A little extra precision in just the right places produces accurate solutions cheaply or demonstrates when problems are too hard to solve without significant cost. This talk will outline the method, iterative refinement with a new twist; the benefits, small backward and forward errors; and the trade-offs and unexpected benefits.","projtag":"lapack, sparse-methods, ieee754","keywords":"linear algebra, sparse matrix, foating point, lapack","ejr-proj":"linear-algebra","bibtex":"@misc{gt09,\n file = {material/gt-2009-08-21.pdf},\n author = {E. Jason Riedy},\n title = {Dependable direct solutions for linear systems using a little extra precision},\n howpublished = {CSE Seminar at Georgia Institute of Technology},\n dom = 21,\n month = aug,\n year = 2009,\n url = {http://hdl.handle.net/1853/29795},\n opttags = {linear algebra; floating point; lapack},\n note = {Invited presentation},\n abstract = {Solving a square linear system $Ax=b$ often is considered a black box. It's supposed to \"just work,\" and failures often are blamed on the original data or subtleties of floating-point. Now that we have an abundance of cheap computations, however, we can do much better. A little extra precision in just the right places produces accurate solutions cheaply or demonstrates when problems are too hard to solve without significant cost. This talk will outline the method, iterative refinement with a new twist; the benefits, small backward and forward errors; and the trade-offs and unexpected benefits.},\n projtag = {lapack, sparse-methods, ieee754},\n keywords = {linear algebra, sparse matrix, foating point, lapack},\n ejr-proj = {linear-algebra}\n}\n\n","author_short":["Riedy, E. J."],"key":"gt09","id":"gt09","bibbaseid":"riedy-dependabledirectsolutionsforlinearsystemsusingalittleextraprecision-2009","role":"author","urls":{"Paper":"http://hdl.handle.net/1853/29795"},"keyword":["linear algebra","sparse matrix","foating point","lapack"],"metadata":{"authorlinks":{}}},"search_terms":["dependable","direct","solutions","linear","systems","using","little","extra","precision","riedy"],"keywords":["linear algebra","sparse matrix","foating point","lapack"],"authorIDs":[],"dataSources":["yxjEg4KYBKLdX7BAz","iovMFeXD2oi65krsd"]}