Subj : Re: A question about atomic_ptr
To   : comp.programming.threads
From : Peter Dimov
Date : Wed Apr 13 2005 07:32 pm

Alexander Terekhov wrote:
> Peter Dimov wrote:
>
> [... weighted reference counting ...]
>
>> The reason I believe that this scheme is not a gain is that it
>> eliminates count increments at the expense of an additional integer
>> in every pointer.
>
> It does bring contention down a bit (at expense of extra space),
> but you still need (naked) atomic operation. Consider: (basic
> thread-safety)
>
>  weighted_reference_counting_ptr w;
>
> Thread A:
>
>  weighted_reference_counting_ptr a(w);
>
> Thread B:
>
>  weighted_reference_counting_ptr b(w);
>
> Both threads need to update w's weight, oder?

Yes (I posted something along these lines in clc++m).

.