• DocumentCode
    1671913
  • Title

    Statistical selection of compiler options

  • Author

    Pinkers, R.P.J. ; Knijnenburg, P.M.W. ; Haneda, M. ; Wijshoff, H.A.G.

  • Author_Institution
    LIACS, Leiden Univ., Netherlands
  • fYear
    2004
  • Firstpage
    494
  • Lastpage
    501
  • Abstract
    Compilers have many switches or options that enable certain code optimizations. However, it is well known that the optimal set of options to be turned on is dependent on both the application and the target architecture. In many cases, standard settings, like -O3, produce suboptimal results due to negative interference of some of the options they contain. We propose an automatic iterative procedure to turn on or to turn off compiler options. This procedure is based on orthogonal arrays that are used for a statistical analysis of profile information to calculate the main effect of the options. We show that our approach outperforms -O3 of GCC on six SPEC benchmarks.
  • Keywords
    iterative methods; optimisation; program compilers; statistical analysis; automatic iterative procedure; code optimization; compiler options; compiler switches; negative interference; orthogonal arrays; statistical analysis; statistical selection; target architecture; Computational modeling; Embedded system; Interference; Optimizing compilers; Program processors; Programming profession; Real time systems; Statistical analysis; Switches; Telecommunication computing;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Modeling, Analysis, and Simulation of Computer and Telecommunications Systems, 2004. (MASCOTS 2004). Proceedings. The IEEE Computer Society's 12th Annual International Symposium on
  • ISSN
    1526-7539
  • Print_ISBN
    0-7695-2251-3
  • Type

    conf

  • DOI
    10.1109/MASCOT.2004.1348305
  • Filename
    1348305