On Tue, 2003-12-16 at 08:14, Phil Payne wrote:
> > I imagined that MVC was "atomic" in nature meaning that the second processor 
> > couldn't
> "garble" things until the MVC had run it's course.
>
> I think that's "implementation-dependent", as IBM would say.  Unwise to rely on it.

Phil is right.  The path between memory and the processor(s) is a twisty
maze of little passages, involving storage alignment, access width,
local caches, interleaving, and Bog knows what else.  Don't rely on
atomic access to *anything* unless the PoOP guarantees it.

In olden days before the compare-and-swap instructions we only had
test-and-set and spin loops.  (And yes, of course, they were coded on
punch cards, carried miles to the computer room in four feet of snow,
uphill, both ways.)

--
David Andrews
A. Duda and Sons, Inc.
[EMAIL PROTECTED]

Reply via email to