Subj : Re: posix and lock-free algorithms
To   : comp.programming.threads
From : David Holmes
Date : Wed Aug 10 2005 03:30 pm

"Joe Seigh" <jseigh_01@xemaps.com> wrote in message
news:8q6dnZ2dnZ0I8bSlnZ2dnVmzb9-dnZ2dRVn-052dnZ0@comcast.com...
> You can see what's Java currently says about this sort of optimization.
> I know an early Java compiler optimized out empty synchronized blocks and
> later had to fix it.  But I don't know if it's allowed with current
> Java semantics.

synchronization on a thread-local object can be completely optimized away
eg:

   synchronized(new Object()) {
       doSomething();
   }

can be transformed to just:

     doSomething();

An empty synchronized block on a non-thread-local object cannot be
completely optimized away, but must preserve the acquire/release semantics
guaranteed by the synchronized block. (In overly simplistic terms, the
locking can be elided but memory barriers might still be needed.)

Cheers,
David Holmes

.