Dynamic optimization


from Wikipedia, the free encyclopedia

In the dynamic optimization is an optimization technique that significantly improves the runtime behavior of software during execution.

Among other things, this makes use of the fact that the values ​​of certain variables in a program are not known before the program is executed, but are constant for a long time while the program is running . So if it is recognized during the execution of a program that a variable seems to be a constant, the program can be compiled as if the variable were actually a constant. This compiled form is then often faster than a compiled form of the program in which a variable is regarded as variable. This compiled form can then run until the value of the variable changes again.

Dynamic optimization is currently only known within virtual machines , since the virtual machine is the one that has to recognize that the value of a variable is constant in order to then recompile the code.

See also

  • HotSpot - Virtual Machine from SUN with dynamic optimization

literature

  • Alfred V. Aho, Ravi Sethi, Jeffrey D. Ullman: Compiler. Principles, Techniques and Tools. ISBN 0-201-10194-7 (The Dragon Book )
  • M. Arnold, SJ Fink, D. Grove, M. Hind, PF Sweeney: A Survey of Adaptive Optimization in Virtual Machines . In: Proceedings of the IEEE . tape 93 , no. 2 , 2005, p. 449-466 , doi : 10.1109 / JPROC.2004.840305 .

Web links