Re: JavaMemoryModel: Why I am resisting putting data/control dependences into the semantics

From: Bill Pugh (pugh@cs.umd.edu)
Date: Sat Aug 11 2001 - 08:29:04 EDT


At 12:12 PM -0400 8/10/01, Doug Lea wrote:
> > I believe that neither data nor control dependences should be used
>> ...
>
>But the alternative is to define "prescient writes" and the like,
>which so far as I understand it, require the same kinds of statements
>("if, under all possible executions...") as I was trying to get
>at. Right?
>
>-Doug

This is a tricky issue.

We avoid the circularities by making assertions about the transformed
program, rather than depending on reasoning (according to some
semantics) about the program before applying a transformation.

        Bill
-------------------------------
JavaMemoryModel mailing list - http://www.cs.umd.edu/~pugh/java/memoryModel



This archive was generated by hypermail 2b29 : Thu Oct 13 2005 - 07:00:34 EDT