The system shall provide dynamic and static information
characterizing the temporal behavior and imposed overhead of any
garbage collection algorithm provided by the system. This
information shall be made available to applications via methods on
GarbageCollector. Implementations are
allowed to provide any set of methods in subclasses as long as the
temporal behavior and overhead are sufficiently categorized. The
implementations are also required to fully document the subclasses.
A reference to the garbage collector responsible for heap memory is
Deprecated. 1.0.1 This class and any subclasses should be singletons.
Preemption latency is a measure of the maximum time a schedulable object may have to wait for the collector to reach a preemption-safe point.
|Methods inherited from class java.lang.Object|
public abstract RelativeTime getPreemptionLatency()
NoHeapRealtimeThread and async event handlers
with the no-heap option
preempt garbage collection immediately, but other schedulable objects
must wait until the collector reaches a preemption-safe point.
For many garbage collectors the only preemption safe point is at the end of
garbage collection, but an implementation of the garbage collector
could permit a schedulable object to preempt
garbage collection before it completes. The
gives such a garbage collector a way to report the worst-case interval between
release of a schedulable object during garbage collection, and the time the schedulable object
starts execution or gains full access to heap memory, whichever comes later.
this. The returned object is allocated in the current allocation context. If there is no constant that bounds garbage collector preemption latency, this method shall return a relative time with
Long.MAX_VALUEmilliseconds. The number of nanoseconds in this special value is unspecified.