Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!swrinde!ucsd!hub.ucsb.edu!appmag!curly!pa
From: pa@curly.appmag.com (Pierre Asselin)
Newsgroups: comp.lang.c
Subject: Re: low level optimization
Message-ID: <703@curly.appmag.com>
Date: 26 Apr 91 01:27:01 GMT
References: <21961@lanl.gov> <11129@ncar.ucar.edu> <22246@lanl.gov> <1991Apr24.174057.22470@ee.eng.ohio-state.edu>
Reply-To: pa@appmag.com (Pierre Asselin)
Organization: Applied Magnetics, Goleta, CA
Lines: 23

It's late.  OK, I'm-a-byt'n...

In article <1991Apr24.174057.22470@ee.eng.ohio-state.edu>
rob@kaa.eng.ohio-state.edu (Rob Carriere) writes:
>In article <22246@lanl.gov> jlg@cochiti.lanl.gov (Jim Giles) spouts forth:
>>P.S.  As I keep pointing out, the _loader_ (or some load-time tool)
>>_can_ satisfy the standard and still do intermodule optimization.
>>The 'translator' (usually thought of as the compiler) cannot.
>
>Err, yes it can.  Someone else has already explained how to do this by
>including optimized and non-optimized versions of the code in the same object
>file. [...]  I would think that 2 implementations count as sufficient
>counterexamples [...]

But does this sheme really count?  Suppose there are N modules subject
to interoptimization.  Translating any one of them leads to a
2^(N-1)-way branch as to what set of optimizations is allowed.
Hmmm...  or 2^M,, where M is the number of optimization tricks the
compiler knows about.  Still too big, though.

On that thought, good night.

  --Pierre Asselin, R&D, Applied Magnetics.  I speak for me.