CLOMP is the C version of the Livermore OpenMP benchmark deeloped to measure OpenMP overheads and other performance impacts due to threading (like NUMA memory layouts, memory contention, cache effects, etc.) in order to influence future system design. Current best-in-class implementations of OpenMP have overheads at least ten times larger than is required by many of our applications for effective use of OpenMP. This benchmark shows the significant negative performance impact of these relatively large overheads and of other thread effects. The CLOMP benchmark highly configurable to allow a variety of problem sizes and threading effects to be studied and it carefully checks its results to catch many common threading errors. This benchmark is expected to be included as part of the Sequoia Benchmark suite for the Sequoia procurement.
To order this software or receive further information, please fill out the following request: Request Software
@misc{osti_1257767,
title = {Clomp, Version 00},
author = {Gylenhaal, J. and Bronevetsky, G.},
abstractNote = {CLOMP is the C version of the Livermore OpenMP benchmark deeloped to measure OpenMP overheads and other performance impacts due to threading (like NUMA memory layouts, memory contention, cache effects, etc.) in order to influence future system design. Current best-in-class implementations of OpenMP have overheads at least ten times larger than is required by many of our applications for effective use of OpenMP. This benchmark shows the significant negative performance impact of these relatively large overheads and of other thread effects. The CLOMP benchmark highly configurable to allow a variety of problem sizes and threading effects to be studied and it carefully checks its results to catch many common threading errors. This benchmark is expected to be included as part of the Sequoia Benchmark suite for the Sequoia procurement.},
doi = {},
url = {https://www.osti.gov/biblio/1257767},
year = {Fri May 25 00:00:00 EDT 2007},
month = {Fri May 25 00:00:00 EDT 2007},
note =
}
Medium: X; OS: This benchmark runs on a wide range of computers that supports OpenMP threading and multiple processors. It has been testeed on IBM, Intel, and AMD platforms at LLNL.; Compatibility: Workstation
Research Organization:
Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)