Re: A Lock-Free vs. Lock-Based Test Here...

> That's not really comparable. Compare (assume no false-sharing)
         ^^^^^^^^^^^^^^^^^^^^^^^

Yup.

>
> atomic_increment(&i);
> atomic_increment(&j);
>
>     to
>
> acquire_fast_lock(l1);
> i++;
> release_fast_lock(l1);
> acquire_fast_lock(l2);
> j++;
> release_fast_lock(l2);

Thats correct sir!

Thanks Alex for pointing that out to David. He seems like he doesn't
believe
anything I say...