ForkJoinPool update

[Mike Duigou]

Soon. I was waiting for http://hg.openjdk.java.net/lambda/lambda/hotspot/rev/ad87adf32248 I will make sure this fixes the encountered problem before adding the F/J patch. I wanted to avoid adding more changes into the mix while our tests were not working. I'll also do another pull from JDK8 mainline.

One important note: The F/J changes import into lambda will be import only. We will not be pushing any of the F/J changes upstream to JDK mainline. ie. JSR166x will flow into OpenJDK the same way as it always has. This may mean that at some point pushes from lambda to mainline might be blocked waiting for JSR166x to be available in mainline.

Mike

On Aug 9 2012, at 08:38 , Aleksey Shipilev wrote:

> Ping? :)
> 
> -Aleksey.
> 
> On 07/31/2012 08:55 PM, Aleksey Shipilev wrote:
>> ForkJoin* is at main tree in JSR166. I had pulled ForkJoin* and
>> CountedCompleter, checked the lambda/lambda build works, and also my
>> microbenchmarks are not failing.
>> 
>> You can pick up the patch and/or changeset here:
>>  http://shipilev.net/pub/jdk/lambda/20120731-fjp-update-1.patch.gz
>>  http://shipilev.net/pub/jdk/lambda/20120731-fjp-update-1.changeset.gz
>> 
>> Thanks!
>> Aleksey.
>> 
>> On 07/31/2012 07:58 PM, Mike Duigou wrote:
>>> If you are willing to generate a changeset merging Doug's jsr166e into lambda I will be very happy to push it.
>>> 
>>> Otherwise we didn't have any specific short term integration plans.
>>> 
>>> Mike
>>> 
>>> On Jul 31 2012, at 06:19 , Aleksey Shipilev wrote:
>>> 
>>>> Hi,
>>>> 
>>>> Does anyone know when the next integration of FJP would occur into JDK
>>>> workspace? My experiments with bulk operations have been stalled on the
>>>> known scalability issues with the old FJP (for one, $submissionLock is
>>>> the scalability bottleneck), which are fixed in new version.
>>>> 
>>>> If there is a hurdle in usual jsr166 -> jdk8 -> lambda integration path,
>>>> can we pull the FJP directly into lambda?
>>>> 
>>>> -Aleksey.
>>>> 
>>> 
>> 
>> 
>