[cpp-threads] Causality on more than two processors, write atomicity

[Doug Lea]

Herb Sutter wrote:
>>
> 
>>And historically, almost no algorithms/programs have ever been
>>found to require such strong guarantees. We once challenged
>>people to come up with non-toy examples, and never got any.
> 
> 
> Really? Off the top of my head (imagine these represent a bounded queue,
> or a series of state changes in a FSM, or a series of barrier/gate
> flags, or...):
> 

You've just listed a bunch of things for which (to the best of
my knowledge) no lock free algorithms exist. (We are still in
the dark ages here.)

I guess what I am saying is that lock-free algorithms have to deal
with so much other chaos that the lack of transitivity of visibility
across observers doesn't seem to make anything any harder. They are
already plenty hard enough though.

-Doug