Issue No.09 - September (2010 vol.43)
Amer Diwan , University of Colorado , Boulder Boulder
Matthias Hauswirth , University of Lugano, Lugano
Peter Sweeney , IBM T.J. Watson Research Center, Hawthorne
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/MC.2010.214
Experiments that do not account for all relevant variables can produce misleading results. The authors identify two variables that compiler optimization evaluations often omit, resulting in significant omitted-variable bias, and discuss how to identify new OVB sources and incorporate them in an experimental design.
Omitted-variable bias, Design and test, Computer performance, Measurement errors, Cache memories
Amer Diwan, Matthias Hauswirth, Peter Sweeney, "The Effect of Omitted-Variable Bias on the Evaluation of Compiler Optimizations", Computer, vol.43, no. 9, pp. 62-67, September 2010, doi:10.1109/MC.2010.214