Multi-threading

TSX anti patterns in lock elision code

Lock elision is a new way to scale programs. It requires following some rules for correctness and good performance. The basic rules are documented in the SDM, chapter 15, and the OPT-GUIDE, chapter 11.

  • Développeurs
  • Professeurs
  • Étudiants
  • Linux*
  • C/C++
  • Avancé
  • TSX
  • Multi-threading
  • Intel® Transactional Synchronization Extensions
  • Optimisation
  • Threading and the Intel® IPP Library – part 3 of 3

    OpenMP Threading and Intel IPP

    The low-level primitives within the IPP library generally represent basic atomic operations. This limits threading within the library to ~15-20% of the primitives. Intel OpenMP is used to implement internal threading and is enabled, by default, when you use one of the multi-threaded variants of the library. Multi-threaded versions of the library are only supported on Linux, Windows, and Mac OS X.

    S’abonner à Multi-threading