DDGVec

From HPCRL Wiki
(Difference between revisions)
Jump to: navigation, search
(Created page with "== Overview ==")
 
(New page for DDGVec)
Line 1: Line 1:
 
== Overview ==
 
== Overview ==
 +
 +
DDGVec is a dynamic trace-based analysis tool used to detect potential SIMD parallelism in the programs that may otherwise be missed by the conservative compile-time analyses. The tool takes an automatic approach to characterize the inherent vectorizability potential of existing applications by analyzing information about run-time dependences and memory access patterns.
 +
 +
== People ==
 +
 +
=== Faculty ===
 +
 +
* [http://www.cse.ohio-state.edu/~saday/ Prof. P. Sadayappan]
 +
* [http://www.cse.ohio-state.edu/~rountev/ Prof. Atanas Rountev]
 +
* [http://www.cs.ucla.edu/~pouchet/ Prof. Louis-Noël Pouchet]
 +
 +
=== Students ===
 +
 +
* [http://jholewinski.org/ Justin Holewinski]
 +
* [http://www.cse.ohio-state.edu/~ramamurr/ Ragavendar Ramamurthi]
 +
* [http://www.cse.ohio-state.edu/~ravishan/ Mahesh Ravishankar]
 +
* [http://www.cse.ohio-state.edu/~fauzia/ Naznin Fauzia]
 +
* [http://www.cse.ohio-state.edu/~elangov/ Venmugil Elango] (Contact person)
 +
 +
== Acknowledgements ==
 +
 +
This software is based upon work supported by the National
 +
Science Foundation under grants CCF-0811781, CCF-0926127, OCI-0904549,
 +
CCF-1017204, and by the Department of Energy's Office of Advanced Scientific
 +
Computing under grant DE-SC0005033. Any opinions, findings, and conclusions
 +
or recommendations expressed in this material are those of the authors and
 +
do not necessarily reflect the views of the National Science Foundation, the
 +
Department of Energy, or The Ohio State University.

Revision as of 16:53, 5 August 2013

Contents

Overview

DDGVec is a dynamic trace-based analysis tool used to detect potential SIMD parallelism in the programs that may otherwise be missed by the conservative compile-time analyses. The tool takes an automatic approach to characterize the inherent vectorizability potential of existing applications by analyzing information about run-time dependences and memory access patterns.

People

Faculty

Students

Acknowledgements

This software is based upon work supported by the National Science Foundation under grants CCF-0811781, CCF-0926127, OCI-0904549, CCF-1017204, and by the Department of Energy's Office of Advanced Scientific Computing under grant DE-SC0005033. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation, the Department of Energy, or The Ohio State University.

Personal tools