Re: Method Level optimizations.

hbaker@netcom.com (Henry Baker)
Fri, 27 Jan 1995 03:54:57 GMT

          From comp.compilers
Related articles
Method Level optimizations. cef@geodesic.com (Charles Fiterman) (1995-01-23)
Re: Method Level optimizations. hbaker@netcom.com (1995-01-27)
Re: Method Level optimizations. monnier@di.epfl.ch (Stefan Monnier) (1995-01-31)
| List of all articles for this month |
Newsgroups: comp.compilers
From: hbaker@netcom.com (Henry Baker)
Keywords: optimize, OOP
Organization: nil
References: 95-01-044
Date: Fri, 27 Jan 1995 03:54:57 GMT

Charles Fiterman <cef@geodesic.com> wrote:


> I'm reaching the conclusion that the best optimizations
> are at very high levels. Consider the optimization of
> A = A + B; to A += B; where A and B are substantial objects
> requiring allocation and initialization.
>
> Consider the advantages of optimizing away A *= unity; Where
> A and unity are matrices. Many of the class definitions I see
> have real opportunities to describe method level optimizations.
> But there are no reflexivity hooks to use them.
>
> This looks like an impressive way to improve performance in
> heavily reused classes.


I agree. Most Lisp compilers have some sort of 'compiler optimizers'
which allow the programmer to give 'hints' such as this to the compiler.
Perhaps it is time for C++ compilers to get their act together.
--

Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.