Mercurial > hg > graal-compiler
comparison src/share/vm/oops/oop.hpp @ 14210:397385cc2c83
8028993: Full collections with ParallelScavenge slower in JDK 8 compared to 7u40
Summary: 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
author | sjohanss |
---|---|
date | Tue, 10 Dec 2013 10:31:00 +0100 |
parents | 740e263c80c6 |
children | cd6b3f1a94ff |
comparison
equal
deleted
inserted
replaced
14209:46e2a5b0c076 | 14210:397385cc2c83 |
---|---|
326 | 326 |
327 // Adjust all pointers in this object to point at it's forwarded location and | 327 // Adjust all pointers in this object to point at it's forwarded location and |
328 // return the size of this oop. This is used by the MarkSweep collector. | 328 // return the size of this oop. This is used by the MarkSweep collector. |
329 int adjust_pointers(); | 329 int adjust_pointers(); |
330 | 330 |
331 #if INCLUDE_ALL_GCS | |
332 // Parallel old | |
333 void update_header(ParCompactionManager* cm); | |
334 #endif // INCLUDE_ALL_GCS | |
335 | |
336 // mark-sweep support | 331 // mark-sweep support |
337 void follow_body(int begin, int end); | 332 void follow_body(int begin, int end); |
338 | 333 |
339 // Fast access to barrier set | 334 // Fast access to barrier set |
340 static BarrierSet* bs() { return _bs; } | 335 static BarrierSet* bs() { return _bs; } |