Differences between the HotSpot in OpenJDK and the closed Oracle JDK

Volker Simonis volker.simonis at gmail.com
Fri Mar 4 03:31:08 PST 2011


Sorry for my impertinence. I already sent this message to the list a
few days before but I can not believe that nobody is interested in
this topic so I resend it with a different (more appropriate?)
subject. If again nobody comments I apologize and promise to keep
silent forevermore:)



I just saw that 6u23 has a new optimization option -XX:+UseCompressedStrings.
See for example:
http://thevirtualmachinist.blogspot.com/2010/12/xxusecompressedstrings-explained.html

Apparently this feature was implemented for the performance release
6u21p as can be seen here:
http://www.oracle.com/technetwork/java/javase/tech/vmoptions-jsp-140102.html

Looking at http://download.java.net/jdk6/6u23/promoted/b01/changes/JDK6u23.list.html
one may guess that this was change:

6836579: Use byte[] instead of char[] for Strings where possible (Hotspot)

but unfortunately the bug report is not available publicly.

6u23 has HS 19b09. If I look at
http://hg.openjdk.java.net/hsx/hsx19/master I can not find this
option, although it is HS19b09 as well.

I thought the general promise was that the HotSpot would be developed
"in the open". I understand that sensitive security fixes may follow a
 "deferred integration" policy, but what about this pure performance
relevant change: it's not in HSX19 neither in the tip revision. Do we
have to expect more proprietary changes which will be present only in
the commercial/closed version of the OpenJDK?

Any comments?

Regards,
Volker


More information about the hotspot-dev mailing list