[concurrency-interest] Performance regression in newest (6/20/2013) jsr166 update

Alex Miller alex at puredanger.com
Wed Jun 26 13:46:14 EDT 2013


Sorry, you're absolutely correct! The algorithmic parallelism in reducers
is the classic fork/join model.


On Wed, Jun 26, 2013 at 12:42 PM, Ron Pressler <ron.pressler at gmail.com>wrote:

>
>
> On Wed, Jun 26, 2013 at 7:31 PM, Alex Miller <alex at puredanger.com> wrote:
>
>>
>> The Clojure reducer library and nascent core.async library are also
>> taking this path so add us to the interested parties watching this use
>> case. :)
>>
>>
> Reducers, too? I would have thought that reducers are a classic
> computation tree, just the kind that would benefit from FJ's default mode,
> which gives you a "depth-first" traversal, which is just what you need if
> you want work-stealing to yield good multi-core load balancing.
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://cs.oswego.edu/pipermail/concurrency-interest/attachments/20130626/635917a1/attachment.html>


More information about the Concurrency-interest mailing list