| 摘要 |
During at least some collection intervals, a garbage collector that employs the train algorithm for collection of a heap generation collects more than one car section. By thus varying the size of the collection set, the collector is able to respond to varying conditions with more flexibility than it could if it collected only a single car section in any interval. When the collector searches for references in portions of a generation identified by an entry in a given car section's remembered set, it searches not only for references to objects in that car section but also to references for objects in other collection-set car sections. This increases scanning efficiency.
|
