Subj : Re: Memory visibility and MS Interlocked instructions
To   : comp.programming.threads
From : Sean Kelly
Date : Thu Sep 01 2005 11:36 pm

Alexander Terekhov wrote:
>
> Memory locations may be replicated (in caches) and the guide
> (mostly toilet paper) doesn't say that "the system guarantees
> atomic propagation of modifications" (or something like that).

Things are becoming more clear as I catch up on the papers that have
been referenced.  However,

http://research.compaq.com/wrl/people/kourosh/papers/1990-rc-isca.pdf

seems to imply that such atomic propogation is a part of the PC model.
But I guess that's assuming Intel's concept of PC is the same :)  I
don't suppose there's enough solid documentation to clear up these
issues?


Sean

.