Related articles |
---|
Information on dispatch-techniques astor@lucifer.guardian.no (Alexander Kjeldaas) (1996-10-30) |
Re: Information on dispatch-techniques roques@pond.sub.org (Christian von Roques) (1996-11-03) |
Re: Information on dispatch-techniques nkramer@cs.cmu.edu (Nick Kramer) (1996-11-05) |
Re: Information on dispatch-techniques volanski@irisa.fr (1996-11-07) |
Re: Information on dispatch-techniques astor@lucifer.guardian.no (Alexander Kjeldaas) (1996-11-26) |
From: | volanski@irisa.fr (Nic Volanski) |
Newsgroups: | comp.compilers |
Date: | 7 Nov 1996 00:09:29 -0500 |
Organization: | IRISA, Campus de Beaulieu, 35042 Rennes Cedex, FRANCE |
References: | 96-11-044 |
Keywords: | OOP |
> (I've seen a lot more papers on caching schemes and on optimizing away
> runtime dispatch than I have seen papers on actually implementing the
> dispatch. Am I just looking in the wrong places?)
> -Nick Kramer
I guess so. Here you have a paper comparing several implementations, via
benchmarks:
H. Srinivassan and P.F. Sweeney, "Evaluating virtual dispatch mechanisms for C++"
Research Report at the IBM T.J. Watson Center - RC 20330 (89860), Nov. 96.
A summary of existing techniques, and a new technique for multi-method
dispathing are covered in:
E. Amiel, O. Gruber and E. Simon, "Optimizing multi-method dispatch using
compressed dispatch tables", appeared in OOPSLA'94.
You can find some further improvement of this technique in:
E. Dujardin, "Efficient dispatch of multi-methods in constant time using
dispatch trees", Research Report at the INRIA, RR 2892, May 96.
Nic.
--
Nic Volanski IRISA - equipe LANDE
volanski@irisa.fr bureau D-056
Fax (+33) 02 99 84 71 71 Tel. 02 99 84 75 86
--
Return to the
comp.compilers page.
Search the
comp.compilers archives again.