RFR (7xS): 8162928: Micro-optimizations in scanning the remembered sets

Thomas Schatzl thomas.schatzl at oracle.com
Fri May 5 11:39:53 UTC 2017


Hi Kim, Mikael,

On Wed, 2017-05-03 at 15:45 +0200, Mikael Gerdin wrote:
> Hi,
> 
> On 2017-04-21 03:47, Kim Barrett wrote:
> > 
> > > 
> > > On Apr 11, 2017, at 7:30 AM, Thomas Schatzl <thomas.schatzl at oracl
> > > e.com> wrote:
> > > 
> > > Hi all,
> > > 
> > >   I would like you to ask for reviews of a set of related changes
> > > that
> > > optimize the way G1 updates/scans/refines cards of the remembered
> > > set.
> > > 
> > > […]
> > > Here we go:
> > > 
> > > […]
> > > 8162928: Micro-optimizations in scanning the remembered sets
> > > 
> > > CR: https://bugs.openjdk.java.net/browse/JDK-8162928
> > > Webrev: http://cr.openjdk.java.net/~tschatzl/8162928/webrev/
> > > 
> > > This change removes the use of HeapRegionDCTOC during scanning
> > > remembered sets, and uses the faster
> > > HeapRegion::oops_on_card_seq_iterate_careful() (measured).
> > > HeapRegionDCTOC also depends CMS code which is just bad.
> > Looks good.
> +1

 thanks for your reviews.

For reference, the webrev I am going to push is as follows, as a change
in the previous CR change the signature of
oops_on_card_seq_iterate_careful:
http://cr.openjdk.java.net/~tschatzl/8162928/webrev.0_to_1/ (diff)
http://cr.openjdk.java.net/~tschatzl/8162928/webrev.1/ (full)

Thanks,
  Thomas



More information about the hotspot-gc-dev mailing list