Sun Microsystems, Inc.

Oracle America, Inc.

Bernd J Mathiske

Oleg A Pliss

Mario I Wolczko

Gregory M Wright

Nedim Fresko

Vijay G Nagarajan

Laurent Daynes

Ioi Kim Lam

Hideya Kawahara

Matthew L Seidl

William R Bush

Kinsley Wong

Nachiappan Periakaruppan

A method for implementing a just-in-time compiler involves obtaining high-level code templates in a high-level programming language, where the high-level programming language is designed for compilation to an intermediate language capable of execution by a virtual machine, and where each high-level code template represents an instruction in the intermediate language. The method further involves compiling the high-level code templates to native code to obtain optimized native code templates, where compiling the high-level code templates is performed, prior to runtime, using an optimizing static compiler designed for runtime use with the virtual machine. The method further involves implementing the just-in-time compiler using the optimized native code templates, where the just-in-time compiler is configured to substitute an optimized native code template when a corresponding instruction in the intermediate language is encountered at runtime.

Method and system for implementing a just-in-time compiler

Techniques for processing native events associated with tasks that are concurrently supported in a virtual machine are disclosed. The techniques allow the user of a virtual machine to concurrently perform tasks even if these tasks require processing of platform-specific (or native) events (e.g., platform-specific input). A smart event-dispatcher is used to dispatch native events to multiple tasks that are concurrently supported by the virtual machine. The smart event-dispatcher can determine which one of a plurality of concurrent tasks should receive the native event for processing. The determination can be made based on a variety of different criteria. In mobile devices, the foreground task is typically the only task that is displayed for the user of mobile devices at any given time. Thus, dispatching events to the foreground task is especially useful for a virtual machine that support a mobile device.

Processing events for concurrent tasks in a virtual machine

Techniques for selectively promoting objects in a generational garbage collector are provided. Rather than using a single promotion policy, information about objects are used to determine and effectively enforce a suitable promotion policy for them. This information, for example, may be information about the general context under which the object is allocated or executed, or it may pertain to a particular operation, thread, function (or method) where garbage is likely to be generated and/or mostly garbage is generated. When it is determined that garbage is likely to be generated or mostly garbage is to be generated, a more suitable promotion policy can be enforced rather than a general (or normal). Objects are allocated using two or more memory allocators. Typically, objects that are likely to be garbage are allocated in a manner that effectively delays or bars their promotion to the next generation.

Selective promotion policy for generational garbage collectors

One embodiment of the present invention provides a memory-management system that supports both address-referenced objects and identifier-referenced objects, wherein an address-referenced object is accessed through a reference containing an address of the object, and wherein an identifier-referenced object is accessed through a reference containing an object identifier (OID) for the object. During operation, the system receives a request to access an object. Next, the system determines if the object is an address-referenced object or an identifier-referenced object. If the object is an address-referenced object, the system accesses the object using the associated address for the object. If the object is an identifier-referenced object, the system accesses the object by using the associated OID for the object to look up the address for the object, and then using the address to access the object.

Memory management system that supports both address-referenced objects and identifier-referenced objects

One embodiment of the present invention provides a system that facilitates faster execution of code on a memory-constrained computing device that has fast on-chip RAM, wherein the fast on-chip RAM is located on a processor chip, but is not cache memory. The system operates by copying a compiled method from an object heap to the fast on-chip RAM on the memory-constrained computing device. Additionally, the system updates an execution pointer to point to the compiled method in the fast on-chip RAM, wherein the execution pointer can also point to a compiled method in the object heap or an interpreted method in the object heap.

Method and apparatus for facilitating faster execution of code on a memory-constrained computing device

One embodiment of the present invention provides a system that supports low-latency session-mobility for an ultra-thin-client. During system operation, an ultra-thin-client sends a location-identifier to a Connection Assignment Server (CAS), which facilitates communication with a user-interface (UI) server, wherein the location-identifier specifies the current location of the ultra-thin-client. Next, the ultra-thin-client receives the address of a local UI-server from the CAS, wherein the CAS selects the local UI-server based on the location-identifier. The ultra-thin-client then sends a user-session identifier to the local UI-server. This allows the local UI-server to retrieve a user-session-image for a user-session from a user-session-image repository. Note that, before moving to the current location, the ultra-thin-client was previously communicating with a remote UI-server, which stored the user-session-image in the user-session-image repository. Next, the ultra-thin-client establishes a user-session connection with the local UI-server. In this way, the system enables the ultra-thin-client to resume the user-session through the local UI-server. Note that the system reduces the delay experienced by the user because it enables the ultra-thin-client to carry on the user-session with the local UI-server instead of the remote UI-server.

Low-latency ultra-thin-client infrastructure

A system and method for reducing the overhead involved in allocating memory to a task, thread or similar entity that shares memory or some other resource with other tasks or threads. A task is assigned a memory limit identifying a maximum amount of memory it may use at one time. When the task requires additional memory to store an object, it is granted a local allocation buffer if the size of the buffer plus the task's current memory allocation will not exceed the task's memory limit. Thereafter, memory space for objects is allocated from the task's local allocation buffer. This scheme avoids the overhead of constantly checking whether it has reached its limit, which is normally performed every time a task allocates memory. For large objects (e.g., greater in size than a local allocation buffer), memory may be allocated directly from the shared area.

System and method for reducing accounting overhead during memory allocation

One embodiment of the present invention provides a system that facilitates keeping track of memory usage of tasks in a shared heap without performing a full garbage-collection operation. The system operates by determining a memory usage by each task in a young generation of the shared heap. Once the memory usage for each task has been determined, the system then adds the memory usage for each task in the young generation of the shared heap to a global memory usage for each task (obtained during a preceding full garbage-collection operation) to produce an actual memory usage for each task.

Method and apparatus for keeping track of memory usage for tasks in a shared heap

A system and method for intelligent generational garbage collection using a dynamic window. During normal memory allocation, a sliding window defines a young generation within an older generation or other area of memory. When data are stored that will become garbage within a finite period of time, a temporary phase of operation is initiated. In the temporary allocation phase, the lower bound of the window is fixed, while the upper bound is allowed to expand to accommodate new objects. When the data become garbage, the window is garbage collected and compacted, and normal memory allocation and garbage collection operations resume. Thus, the window is dynamic in both movement and size. When the temporary allocation phase is initiated, the young generation may be garbage collected and compacted, and the lower window bound may be fixed at the location (e.g., address) where the allocation point was when the target data were stored.

Growing community of inventors

Cupertino, CA, United States of America

Bernd J Mathiske