RFC (M) 8058968: Compiler time traces should be improved

John Rose john.r.rose at oracle.com
Wed Sep 24 22:40:57 UTC 2014


On Sep 24, 2014, at 1:09 PM, Aleksey Shipilev <aleksey.shipilev at oracle.com> wrote:

> I think going forward we should convert the entire thing into JFR
> counters, and let the data parsers outside JVM handle all the
> aggregation and statistics. If you look in the compilerBroker time
> tracing code, it is an large slab of tracing boiler-plate...

Yes, that's better; then we can have histos etc.
JFR can handle that volume of data points easily.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/hotspot-compiler-dev/attachments/20140924/8d66971d/attachment-0001.html>


More information about the hotspot-compiler-dev mailing list