Request for review (S): 8005396: Use ParNew with only one thread instead of DefNew as default for CMS on single CPU machines

Jon Masamitsu jon.masamitsu at oracle.com
Fri Dec 21 14:43:20 PST 2012


http://cr.openjdk.java.net/~brutisso/8005396/webrev.00/src/share/vm/runtime/arguments.cpp.frames.html

1098       FLAG_SET_DEFAULT(ParallelGCThreads, par_gc_threads);

I'd suggest using FLAG_SET_ERGO instead of FLAG_SET_DEFAULT.



http://cr.openjdk.java.net/~brutisso/8005396/webrev.00/src/share/vm/gc_implementation/parNew/parNewGeneration.cpp.frames.html

1627 bool ParNewGeneration::in_use() {
1628   return UseParNewGC;
1629 }

Is "in_use()" useful anymore?  It was invented to avoid a direct
use of ParallelGCThreads.   Maybe just replace it with UseParNewGC?

The rests looks good.

Jon

On 12/21/2012 5:42 AM, Bengt Rutisson wrote:
>
> Hi All,
>
> Can I have a couple of reviews for this change?
>
> http://cr.openjdk.java.net/~brutisso/8005396/webrev.00/
>
> Currently we use ParNew as default for the young generation when CMS 
> is selected. But if the machine only has a single CPU we set the 
> ParallelGCThreads to 0 and and select DefNew instead of ParNew.
>
> As part of another change, 8003820, we will deprecate the DefNew + CMS 
> combination. Thus, it does not make sense anymore to have this 
> selected by default. This fix is to make CMS always pick ParNew by 
> default.
>
> The change also has the side effect that the, in my opinion, rather 
> strange behavior that setting ParallelGCThreads=0 on the command line 
> overrides the GC choice. I would expect this command line to give me 
> ParNew, but it actually gives me DefNew:
>
> -XX:+UseParNewGC -XX:ParallelGCThreads=0
>
> After my proposed change you get ParNew with the above command line.
>
> I have done some performance testing to verify that ParNew with one 
> thread is not slower than DefNew. The details are in the bug report:
>
> http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=8005396
>
> but as a summary it can be said that there is no noticeable difference.
>
> I am also running some more SPECjbb2005 runs and will analyze the gc 
> times.
>
> Thanks,
> Bengt


More information about the hotspot-gc-dev mailing list