[concurrency-interest] Array allocation and access on the JVM

Aleksandar Prokopec aleksandar.prokopec at gmail.com
Thu Jan 26 07:46:43 EST 2012


If I understood correctly, this marking is not due to tracking if there 
exists a reference from the tenured to the young generation in the card.

It's an optimization in the young collector which allows completely 
skipping the scan of some cards during GC if there have been no writes 
in them.

Cheers,
Aleksandar


On 1/26/12 1:32 PM, concurrency-interest-request at cs.oswego.edu wrote:
> Message: 1
> Date: Wed, 25 Jan 2012 19:22:48 -0500
> From: Vitaly Davidovich<vitalyd at gmail.com>
> To: Doug Lea<dl at cs.oswego.edu>
> Cc: concurrency-interest at cs.oswego.edu
> Subject: Re: [concurrency-interest] Array allocation and access on the
> 	JVM
> Message-ID:
> 	<CAHjP37EBq=rnu7Xb6yrgpYEnFZ0JGGarxctJjsWcenHTSwF76w at mail.gmail.com>
> Content-Type: text/plain; charset="iso-8859-1"
>
> Slightly off topic but does the VM even need to mark cards when the
> references are both in young gen? I thought they were needed for cases
> where a tenured object references a young gen object; otherwise it seems
> like scanning from the roots is sufficient? I assume the cost of figuring
> out whether it's a cross generation write is too prohibitive to do for
> every pointer store and it just always marks.
>
> Sent from my phone
> On Jan 25, 2012 7:13 PM, "Vitaly Davidovich"<vitalyd at gmail.com>  wrote:
>
>



More information about the Concurrency-interest mailing list