[OpenJDK Rasterizer] Marlin renderer contribution for review

[Laurent Bourgès]

Jim,

thanks for your advice, I agree the webrev is now against pisces:

http://cr.openjdk.java.net/~lbourges/marlin/marlin.1/


I did not check if pisces evolved since march 2013 !

Could someone give me the pisces changelog (from mercurial) ?


Here is the performance summary:

http://cr.openjdk.java.net/~lbourges/marlin/report/perf_summary.htm


*Threads* *4* *1* *2* *4*







*Renderer* *Test count* *27* *9* *9* *9*







*ductus* *Pct95* 355,442 157,616 263,145 645,566







*pisces* *Pct95* 398,576 345,619 373,08 477,03







*Marlin 0.5.6* *Pct95* 130,718 129,718 130,557 131,877







*Marlin JDK9* *Pct95* 137,709 134,898 138,736 139,494


All details (benchmark log files) are published:

http://cr.openjdk.java.net/~lbourges/marlin/report/


PS: do you know if hotspot or JVM flags changed a lot that could explain
the small slowdown with OpenJDK9 ?

I may check JVM flags using -XX:+PrintCommandLineFlags -XX:+PrintFlagsFinal
...


Cheers,

Laurent
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/graphics-rasterizer-dev/attachments/20150313/42e4b7df/attachment.html>