[concurrency-interest] Volatile field use cases?

Kevin Condon conivek at gmail.com
Fri Oct 27 08:03:55 EDT 2006


I don't have any unique uses for volatiles to describe, but I think
this is an important topic.  When I started using Java back around
1999, the rumor was that volatile semantics were not reliable in all
JVM implementations and that the use of synchronized blocks was the
approach of choice.  The rumors were not incredibly well
substantiated, but it was enough that many of us put volatiles in the
closet.  That's a habit that I'm beginning to change with the new JMM,
and this kind of discussion could help to stimulate that change.

Regards,
Kevin

On 10/27/06, Jeremy Manson <jmanson at cs.umd.edu> wrote:
> Hi folks,
>
> Some of us were talking off line about the ways in which people use
> volatile fields.  They can be used for writing a variety of interesting
> data structures, like non-blocking queues, but I have no idea how many
> people are doing this.
>
> I thought it might be interesting to ask the crowd here about the more
> interesting ways in which you use volatiles.  Do you have something
> fabulously interesting that you do?  Do you write non-blocking data
> structures?  One thing I don't really know is if people ever write code
> that relies on strong guarantees between more than one volatile field --
> do you?
>
> Basically, what are you folks using them to do?
>
> All of this is by way of finding out what the impact of the JMM has
> been, now that we are a couple of years down the line: seeing what works
> and what doesn't, and laying out plans for future memory models.
>
> Thanks,
>
>                                        Jeremy
> _______________________________________________
> Concurrency-interest mailing list
> Concurrency-interest at altair.cs.oswego.edu
> http://altair.cs.oswego.edu/mailman/listinfo/concurrency-interest
>


More information about the Concurrency-interest mailing list