JEP 189: Shenandoah: An Ultra-Low-Pause-Time Garbage Collector

[Krystal Mok]

Hi Roman and Christine,

Looking forward to your FOSDEM talk, too.

I'm wondering if you're going to scan all thread stacks in a global
safepoint, or are you going to use a "on-the-fly" style?

Also, it seems like you're not heading in a generational direction. G1
started out without logical generations, but then added a generational mode
and made it the default mode.
Is there any implications on the choice of remember sets and the choice of
not doing generational?

Best regards,
Kris


On Fri, Jan 17, 2014 at 1:18 PM, Monica Beckwith <monica.b at servergy.com>wrote:

> +1 (on FOSDEM talk and Volker's question). And I can help with testing on
> PPC as well.
>
> -Monica
>
>
> On 1/17/14, 3:09 PM, Volker Simonis wrote:
>
>> Hi Roman, Christine,
>>
>> I'm really looking forward to hear your talk about Shenandoah on FOSDEM.
>>
>> But now that you already asked for questions:
>>
>> Will you design the new collector with weak memory model architectures
>> in mind. I know that it is hard if you don't have the corresponding
>> hardware but I think RedHat should at least have access to some decent
>> PowerPC machines and I would strongly advise you to test there
>> thoroughly from the very beginning. We had (and still have) a lot of
>> problems with G1 on these architectures.
>>
>> Regards,
>> Volker
>>
>> On Fri, Jan 17, 2014 at 8:28 PM, Roman Kennke <rkennke at redhat.com> wrote:
>>
>>> Hi there,
>>>
>>> Sorry being so silent... we are very happy that the Shenandoah JEP has
>>> finally been posted and would welcome any comments, questions,
>>> discussion from the community.
>>>
>>> Cheers,
>>> Roman
>>>
>>> Am Mittwoch, den 15.01.2014, 11:49 -0800 schrieb
>>> mark.reinhold at oracle.com:
>>>
>>>> Posted: http://openjdk.java.net/jeps/189
>>>>
>>>> - Mark
>>>>
>>>
>>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/hotspot-gc-dev/attachments/20140117/211ad31b/attachment.htm>