JEP 173: Remove Rarely-Used Combinations of Garbage Collectors
bengt.rutisson at oracle.com
Thu Dec 6 05:32:32 PST 2012
Thanks for your feedback!
On 12/6/12 1:23 AM, Peter B. Kessler wrote:
> mark.reinhold at oracle.com wrote:
>> Posted: http://openjdk.java.net/jeps/173
>> - Mark
> If the only use of SerialOld will be with ParallelScavenge +
> SerialOld, then maybe we should check what the performance hit (GC
> time, heap utilization, application throughput) is using
> ParallelScavenge + ParallelOld, possibly with ParallelOld throttled
> down to 1 worker thread. (I understand that we can't remove the code
> for SerialOld because it backs CMS, but we could improve the user
> experience by not letting people choose sub-optimal configurations.)
> Similarly, if we could start weaning users off of DefNew we might
> eventually be able to remove the code for it.
These are very good observations. In fact I look at the currently
proposed JEP as a first step in a direction towards much fewer GC
combinations. I think what you suggest is a good path forward, but there
may be other step we want to take as well.
> It would help the write-up if the *supported* configurations were listed.
Good point. I'll try to write that up and add it to the JEP.
> ... peter
More information about the hotspot-gc-dev