Abstract: A memory allocation scheme for object allocation in concurrent object-oriented server-side environments that combines the time efficiency of thread-local allocation schemes with the space efficiency of shared-heap allocation schemes has been developed. A pool of memory allocation spaces is maintained so that when a thread transitions to a runnable state, it is allocated a private memory space. However, when the thread transitions out of the runnable state and no longer needs its allocated memory, the space is re-allocated back to the pool for use by another runnable state thread.

Abstract: Methods and apparatus for speculatively locking an object are disclosed. According to one aspect of the present invention, a method for acquiring use of an object using a current thread includes a determination of whether a first bit included in the object is set to indicate that the object is speculatively owned by a speculative owner thread. When the object is speculatively owned, the speculative owner thread is allowed to use the object without locking the object. The method also includes checking a stored identifier that is associated with the object and identifies the speculative owner thread, as well as determining whether the stored identifier identifies the current thread. When the stored identifier identifies the current thread, the current thread already has use of the object; i.e., the current thread is the speculative owner thread.

Abstract: Methods and apparatus consistent with the invention provide improved organization of documents responsive to a search query. In one embodiment, a search query is received and a list of responsive documents is identified. The responsive documents are organized based in whole or in part on usage statistics.