Related articles |
---|
Languages that give "hints" to the compiler jhummel@cy4.ICS.UCI.EDU (Joe Hummel) (1994-07-18) |
Re: Languages that give "hints" to the compiler wisej@acf4.nyu.edu (1994-07-29) |
Re: Languages that give "hints" to the compiler tgl@netcom.com (1994-07-19) |
Re: Languages that give "hints" to the compiler jfisher@hplabsz.hpl.hp.com (1994-07-21) |
Re: Languages that give "hints" to the compiler rfg@netcom.com (1994-07-31) |
Languages that give "hints" to the compiler ssimmons@convex.com (1994-08-01) |
Re: Languages that give "hints" to the compiler jls@summit.novell.com (1994-08-02) |
Re: Languages that give "hints" to the compiler rfg@netcom.com (1994-08-03) |
Newsgroups: | comp.compilers |
From: | tgl@netcom.com (Tom Lane) |
Keywords: | optimize |
Organization: | Netcom Online Communications Services |
References: | 94-07-066 |
Date: | Tue, 19 Jul 1994 06:07:19 GMT |
I believe HP's current PA-RISC compilers support the "scientific" approach
to branch optimization: you compile your program, you run the profiler to
*find out* which branches are most often taken, then you recompile with
the profiler results automatically taken into account.
But I don't have the details at hand. Perhaps someone at HP can fill us
in.
regards, tom lane
PS: obviously this approach is only as good as the test cases you use for
profiling. Still, we all know that that beats programmer intuition, which
is usually the data source for other methods.
--
Return to the
comp.compilers page.
Search the
comp.compilers archives again.