Review request for JDK-8141446: Cache Class.forName for permanently loaded classes
sundararajan.athijegannathan at oracle.com
Thu Nov 5 10:58:57 UTC 2015
Class.forName itself should check classloader's cache, right? Strange
that our own cache on top of it improves perf. but, if we get better
numbers, why not :) ?
On 11/4/2015 10:28 PM, Attila Szegedi wrote:
> Please review JDK-8141446 "Cache Class.forName for permanently loaded classes" at <http://cr.openjdk.java.net/~attila/8141446/webrev.jdk9> for <https://bugs.openjdk.java.net/browse/JDK-8141446>
> This was obtained by profiling the cost of compile-octane.js test, and has an overall benefit on “ant test”. On my machine “ant test” went from 12m30s to 10m30s.
More information about the nashorn-dev