Summary: Searching for the Sorting Record: Experiences in Tuning NOW­Sort
Andrea C. Arpaci­Dusseau, Remzi H. Arpaci­Dusseau,
David E. Culler, Joseph M. Hellerstein, and David A. Patterson
Computer Science Division,
University of California, Berkeley
fdusseau,remzi,culler,jmh,pattrsng@cs.berkeley.edu
Abstract
We present our experiences in developing and tuning the perfor­
mance of NOW­Sort, a parallel, disk­to­disk sorting algorithm.
NOW­Sort currently holds two world records in database­
industry standard benchmarks. Critical to the tuning process
was the setting of expectations, which tell the programmer both
where to tune and when to stop. We found three categories of
useful tools: tools that help set expectations and configure
the application to different hardware parameters, visualization
tools that animate performance counters, and search tools that
track down performance anomalies. All such tools must in­
teract well with all layers of the underlying software (e.g., the
operating system), as well as with applications that leverage
modern OS features, such as threads and memory­mapped I/O.

Source: Arpaci-Dusseau, Andrea - Department of Computer Sciences, University of Wisconsin at Madison
Arpaci-Dusseau, Remzi - Department of Computer Sciences, Department of Computer Sciences, University of Wisconsin at Madison
Culler, David E. - Department of Electrical Engineering and Computer Sciences, University of California at Berkeley
Hellerstein, Joseph M. - Department of Electrical Engineering and Computer Sciences, University of California at Berkeley

Collections: Computer Technologies and Information Sciences