8017163: minor question regarding the calculation and logging of card set coarsening statistics

[Tianqi Xia]

Sure thing, I will take a look.

Thanks,
Tianqi

On Tue, May 17, 2022 at 11:42 PM Thomas Schatzl <thomas.schatzl at oracle.com>
wrote:

> Hi,
>
> On 17.05.22 16:38, Tianqi Xia wrote:
> > Hi,
> >
> > I continued to play around with JDK-8017163: G1: Refactor remembered
> sets,
> > after going through the code, I was really amazed by the delicate magic
> of
> > all those different card containers.
> >
> > I ran BigRamTester with a 20GB heap for a couple minutes, observed the
> logs
> > of coarsening statistics, and noticed the data displayed after
> "Coarsening
> > (recent)" seems a little bit abnormal, here is what I mean:
> >
>
> [...]
>
> >  From the comments in code, "Coarsening (recent)" represents the number
> of
> > coarsening happened during the lastest GC cycle, in my understanding it
> > should be calculated by something like:
> >
> > current total coarsening - previous total coarsening
> >
> > However, after looking at G1CardSet::print_coarsen_stats) the recent
> > coarsening stats are instead calculated by:
> >
> > current total coarsening - coarsening happened in previous(the one before
> > the latest) GC cycle
> >
> > Am I missing something here?
> >
>
> I filed https://bugs.openjdk.java.net/browse/JDK-8286893. Maybe you have
> time to look into this?
>
> Thanks,
>    Thomas
>