Review + pull request: update FJP
aleksey.shipilev at oracle.com
Tue Oct 30 08:38:01 PDT 2012
It turns out, there's the final version of the new update to FJP coming
to jsr166. Please hold on while I'm updating the patch.
On 10/30/2012 03:44 PM, Aleksey Shipilev wrote:
> Sure, but I think lambda/lambda is already ahead, and for a reason:
> performance investigations for lambdas require latest and greatest code.
> In this particular case, this involves more changes in lambda code itself.
> The jsr166 -> jdk8 -> lambda turnaround is longer than bearable for
> active development.
> On 10/30/2012 03:41 PM, David Holmes wrote:
>> There are plans to sync up jdk8 with latest jsr166 changes. Once that
>> happens they can appear in the lambda repos too.
>> On 30/10/2012 8:50 PM, Aleksey Shipilev wrote:
>>> Please consider pulling up the update for new ForkJoinPool from jsr166.
>>> Doug had implemented a few features which are to provide much better
>>> wakeup performance. (Brian, this is the same patch I've cc'ed you
>>> Here's the patch against JDK:
>>> - custom builds on Linux x86_64;
>>> - decomposition performance benchmarks;
>>> - fjp-trace logs
>>> I could generate the webrev, if that is more appropriate.
More information about the lambda-dev