Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!hplabs!amdcad!crackle!tim
From: tim@crackle.amd.com (Tim Olson)
Newsgroups: comp.lang.c
Subject: Re: C optimizer
Message-ID: <24433@amdcad.AMD.COM>
Date: 13 Feb 89 20:10:02 GMT
References: <515@larry.UUCP>
Sender: news@amdcad.AMD.COM
Reply-To: tim@amd.com (Tim Olson)
Organization: Advanced Micro Devices, Inc. Sunnyvale CA
Lines: 28
Summary:
Expires:
Sender:
Followup-To:

In article <515@larry.UUCP> jwp@larry.UUCP (Jeffrey W Percival) writes:
| 
| 	x = (1 + cos(r)) / (cos(r) * sin(r));
| 	y = (cos(r) - sin(r)) / (1 + sin(r));
| 
| I made this up, but the point is the re-use of the sin() and cos()
| calls.  Now, can I expect the compiler to form only one call to sin and
| cos?  Specifically, I am using ULTRIX 2.1 on a VS2000, but my real
| question is, would *most* compilers optimize this as I'd hope, or
| *some* compilers, or hardly any?

I think that very few compilers currently perform this optimization. 
The problem is that (pre-ANSI) compilers usually have no knowledge about
standard library routines -- they look just like other function calls. 
Therefore, the compiler does not know that sin() or cos() are "true
functions", in that they will return the same values for the same
inputs, and have no side-effects.

ANSI-conforming C compilers will be able to perform this optimization,
since the standard library routine names are reserved, and the compiler
can have specific knowledge about them.

I know of at least one C compiler that is tracking the proposed ANSI
standard that will perform this optimization.

	-- Tim Olson
	Advanced Micro Devices
	(tim@crackle.amd.com)