commit | 5f4d0e9c253b8d9c76b5dd1fb6fc5189cf0e254a | [log] [tgz] |
---|---|---|
author | Stefan Johansson <sjohanss@openjdk.org> | Tue Dec 10 10:31:00 2013 +0100 |
committer | Stefan Johansson <sjohanss@openjdk.org> | Tue Dec 10 10:31:00 2013 +0100 |
tree | 320207059baa30bea372d1461a944587790de0b7 | |
parent | 5d5b1157d16ce0163a4172f32cfd2bb69978a5ab [diff] |
8028993: Full collections with ParallelScavenge slower in JDK 8 compared to 7u40 Reducing the number of calls to follow_class_loader to speed up the marking phase. Also removed some unnecessary calls to adjust_klass. Reviewed-by: stefank, jmasa, mgerdin