SC12 Home > SC12 Schedule > SC12 Presentation - Build to Order Linear Algebra Kernels

SCHEDULE: NOV 10-16, 2012

When viewing the Technical Program schedule, on the far righthand side is a column labeled "PLANNER." Use this planner to build your own schedule. Once you select an event and want to add it to your personal schedule, just click on the calendar icon of your choice (outlook calendar, ical calendar or google calendar) and that event will be stored there. As you select events in this manner, you will have your own schedule to guide you through the week.

Build to Order Linear Algebra Kernels

SESSION: Research Poster Reception

EVENT TYPE: Posters and Electronic Posters

TIME: 5:15PM - 7:00PM

SESSION CHAIR: Torsten Hoefler

AUTHOR(S):Christopher Gropp, Geoffrey Belter, Elizabeth Jessup, Thomas Nelson, Boyana Norris, Jeremy Siek

ROOM:East Entrance

ABSTRACT:
Tuning linear algebra kernels for specific machines is a time-consuming and difficult process. The Build-to-Order (BTO) compiler takes MATLAB-like kernels and generates fully tuned C code for the machine on which it runs. BTO allows users to create optimized linear algebra kernels customized to their needs and machines. Previous work has shown that BTO kernels perform competitively with or better than hand-tuned code. We now test BTO on a full application. We have selected bidiagonalization, a non-trivial matrix operation useful in computing singular value decompositions. We use BTO to tune the core kernels of the algorithm, rather than the entire application. This poster shows the comparative performance of four implementations of bidiagonalization: the LAPACK routine DGEBD2, a BLAS 2.5 based algorithm developed by Howell et al., and those same making use of BTO kernels.

Chair/Author Details:

Torsten Hoefler (Chair) - ETH Zurich

Christopher Gropp - Rose-Hulman Institute of Technology

Geoffrey Belter - University of Colorado Boulder

Elizabeth Jessup - University of Colorado Boulder

Thomas Nelson - University of Colorado Boulder

Boyana Norris - Argonne National Laboratory

Jeremy Siek - University of Colorado Boulder

Add to iCal  Click here to download .ics calendar file

Add to Outlook  Click here to download .vcs calendar file

Add to Google Calendarss  Click here to add event to your Google Calendar

Build to Order Linear Algebra Kernels

SESSION: Research Poster Reception

EVENT TYPE:

TIME: 5:15PM - 7:00PM

SESSION CHAIR: Torsten Hoefler

AUTHOR(S):Christopher Gropp, Geoffrey Belter, Elizabeth Jessup, Thomas Nelson, Boyana Norris, Jeremy Siek

ROOM:East Entrance

ABSTRACT:
Tuning linear algebra kernels for specific machines is a time-consuming and difficult process. The Build-to-Order (BTO) compiler takes MATLAB-like kernels and generates fully tuned C code for the machine on which it runs. BTO allows users to create optimized linear algebra kernels customized to their needs and machines. Previous work has shown that BTO kernels perform competitively with or better than hand-tuned code. We now test BTO on a full application. We have selected bidiagonalization, a non-trivial matrix operation useful in computing singular value decompositions. We use BTO to tune the core kernels of the algorithm, rather than the entire application. This poster shows the comparative performance of four implementations of bidiagonalization: the LAPACK routine DGEBD2, a BLAS 2.5 based algorithm developed by Howell et al., and those same making use of BTO kernels.

Chair/Author Details:

Torsten Hoefler (Chair) - ETH Zurich

Christopher Gropp - Rose-Hulman Institute of Technology

Geoffrey Belter - University of Colorado Boulder

Elizabeth Jessup - University of Colorado Boulder

Thomas Nelson - University of Colorado Boulder

Boyana Norris - Argonne National Laboratory

Jeremy Siek - University of Colorado Boulder

Add to iCal  Click here to download .ics calendar file

Add to Outlook  Click here to download .vcs calendar file

Add to Google Calendarss  Click here to add event to your Google Calendar