RFR (S) 8225229: Shenandoah: trim down default number of GC threads

Roman Kennke rkennke at redhat.com
Tue Jun 4 18:41:32 UTC 2019


That looks good to me! Yes, I also always trimmed GC threads ~like this.

Roman

> RFE:
>   https://bugs.openjdk.java.net/browse/JDK-8225229
> 
> We can trim down the default number of GC threads to make Shenandoah more cooperative with the
> application and OS scheduler. We routinely trim down concurrent threads when deploying, and better
> default would make deployment stories simpler. Various tests show that defaulting to 1/4*N_CPU
> concurrent and 1/2*N_CPU parallel threads hits the sweet spot. New test verifies boundary conditions.
> 
> Patch:
>   http://cr.openjdk.java.net/~shade/8225229/webrev.01/
> 
> Testing: hotspot_gc_shenandoah {fastdebug|release}
> 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: OpenPGP digital signature
URL: <https://mail.openjdk.java.net/pipermail/hotspot-gc-dev/attachments/20190604/3cd6bfec/signature.asc>


More information about the hotspot-gc-dev mailing list