Path: utzoo!utgpu!water!watmath!clyde!rutgers!rochester!cornell!uw-beaver!mit-eddie!husc6!hao!noao!arizona!mike
From: mike@arizona.edu (Mike Coffin)
Newsgroups: comp.lang.c
Subject: Re: More on noalias
Summary: It can already be done
Message-ID: <3433@megaron.arizona.edu>
Date: 12 Jan 88 20:34:26 GMT
References: <10129@mimsy.UUCP>
Organization: U of Arizona CS Dept, Tucson
Lines: 38

In article <10129@mimsy.UUCP>, chris@mimsy.UUCP (Chris Torek) writes:
> First, let me construct an example in which aliasing is relatively
> hard to detect.  How about this:
> 
> 	void
> 	vector_add /* maybe */ (int *a, int i, int j, int k, int n)
> 	{
> 
> 		for (; --n >= 0; i++, j++, k++)
> 			a[i] = a[j] + a[k];
> 	}
> 
> This can use a vector add iff i != j and i != k (on some hypothetical
> machine).  Hence the compiler wants to know whether this will be
> true.  By declaring `int *noalias a' or `noalias int *a' (whichever
> it is), we could say that.

I wish I knew for sure what noalias meant.  In any case, you can already
(in good old K&R C) write the above function in a way that allows vector
add:

 	void
 	vector_add /* maybe */ (int *a, int i, int j, int k, int n)
 	{
 		if (i==j || i==k) return;

		/* now the compiler can infer that i!=j and i!=k */
 		for (; --n >= 0; i++, j++, k++)
 			a[i] = a[j] + a[k];
 	}

If you don't like the way the if-statement looks, embed it in a macro;
e.g., "opt_assert(i!=j && i !=k)".
-- 

Mike Coffin				mike@arizona.edu
Univ. of Ariz. Dept. of Comp. Sci.	{allegra,cmcl2,ihnp4}!arizona!mike
Tucson, AZ  85721			(602)621-4252