[concurrency-interest] ForkJoin updates
dl at cs.oswego.edu
Mon Feb 20 13:40:33 EST 2012
On 02/03/12 13:28, Doug Lea wrote:
> On 02/03/12 11:29, Jason Mehrens wrote:
>> Would unconditionally marking the ForkJoinTask as ineligible for biased locking,
> I don't plan to do this, but resolving this in a better way is
> among the "other" remaining changes for this round I mentioned.
Which are now committed, and ...
> The unluckiest of these cases cascade into
> stalls involving many threads/cores, which are the main cause
> of the unwanted positive feedback loops I've mentioned that
> force compensation to be overly conservative.
... which enabled a less conservative compensation
policy, giving some further speedups in some use cases.
These updates also include re-implementing timed gets
under new scheme rather than relying on some
placeholder code that could mis-report status, so if you
are using recent versions, you should get an update.
Available in the usual places -- see
More information about the Concurrency-interest