Abstract: A method and apparatus for providing optimized strong atomicity operations for non-transactional writes is herein described. Locks are acquired upon initial non-transactional writes to memory locations. The locks are maintained until an event is detected resulting in the release of the locks. As a result, in the intermediary period between acquiring and releasing the locks, any subsequent writes to memory locations that are locked are accelerated through non-execution of lock acquire operations.

Abstract: In one embodiment, the present invention includes a method for selecting a first transaction execution mode to begin a first transaction in a unbounded transactional memory (UTM) system having a plurality of transaction execution modes. These transaction execution modes include hardware modes to execute within a cache memory of a processor, a hardware assisted mode to execute using transactional hardware of the processor and a software buffer, and a software transactional memory (STM) mode to execute without the transactional hardware. The first transaction execution mode can be selected to be a highest performant of the hardware modes if no pending transaction is executing in the STM mode, otherwise a lower performant mode can be selected. Other embodiments are described and claimed.

Abstract: Minimizing code duplication in an unbounded transactional memory system. A computing apparatus including one or more processors in which it is possible to use a set of common mode-agnostic TM barrier sequences that runs on legacy ISA and extended ISA processors, and that employs hardware filter indicators (when available) to filter redundant applications of TM barriers, and that enables a compiled binary representation of the subject code to run correctly in any of the currently implemented set of transactional memory execution modes, including running the code outside of a transaction, and that enables the same compiled binary to continue to work with future TM implementations which may introduce as yet unknown future TM execution modes.

Abstract: A method and system to facilitate a user level application executing in a first processing unit to enqueue work or task(s) safely for a second processing unit without performing any ring transition. For example, in one embodiment of the invention, the first processing unit executes one or more user level applications, where each user level application has a task to be offloaded to a second processing unit. The first processing unit signals the second processing unit to handle the task from each user level application without performing any ring transition in one embodiment of the invention.

Abstract: In one embodiment, the invention provides a method comprising determining metadata encoded in instructions of a stored program; and executing the stored program based on the metadata.

Abstract: A method and apparatus for accelerating a software transactional memory (STM) system is described herein. Annotation field are associated with lines of a transactional memory. An annotation field associated with a line of the transaction memory is initialized to a first value upon starting a transaction. In response to encountering a read operation in the transaction, then annotation field is checked. If the annotation field includes a first value, the read is serviced from the line of the transaction memory without having to search an additional write space. A second and third value in the annotation field potentially indicates whether a read operation missed the transactional memory or a tentative value is stored in a write space. Additionally, an additional bit in the annotation field, may be utilized to indicate whether previous read operations have been logged, allowing for subsequent redundant read logging to be reduced.

Abstract: In one embodiment, the present invention includes a method for executing a transactional memory (TM) transaction in a first thread, buffering a block of data in a first buffer of a cache memory of a processor, and acquiring a write monitor on the block to obtain ownership of the block at an encounter time in which data at a location of the block in the first buffer is updated. Other embodiments are described and claimed.

Abstract: Operating system virtual memory management for hardware transactional memory. A system includes an operating system deciding to unmap a first virtual page. As a result, the operating system removes the mapping of the first virtual page to the first physical page from the virtual memory page table. As a result, the operating system performs an action to discard transactional memory hardware state for at least the first physical page. Embodiments may further suspend hardware transactions in kernel mode. Embodiments may further perform soft page fault handling without aborting a hardware transaction, resuming the hardware transaction upon return to user mode, and even successfully committing the hardware transaction.

Abstract: A example method disclosed herein comprises initiating a first optimistically balanced synchronization to acquire a lock of an object, the first optimistically balanced synchronization comprising a first optimistically balanced acquisition and a first optimistically balanced release to be performed on the lock by a same thread and at a same nesting level, releasing the lock after execution of program code covered by the lock if a stored state of the first optimistically balanced release indicates that the first optimistically balanced release is still valid, the stored state of the first optimistically balanced release being initialized prior to execution of the program code to indicate that the first optimistically balanced release is valid, and throwing an exception after execution of the program code covered by the lock if the stored state of the first optimistically balanced release indicates that the first optimistically balanced release is no longer valid.

Abstract: Methods for mapping a function pointer to the device code are presented. In one embodiment, a method includes identifying a function which is executable by processing devices. The method includes generating codes including a first code corresponds to a first processing device and a second code corresponds to a second processing device. The second processing device is architecturally different from the first processing device. The method further includes storing the second code in a byte string such that the second code is retrievable if the function will be executed by the second processing device.

Abstract: A method and apparatus to facilitate shared pointers in a heterogeneous platform. In one embodiment of the invention, the heterogeneous or non-homogeneous platform includes, but is not limited to, a central processing core or unit, a graphics processing core or unit, a digital signal processor, an interface module, and any other form of processing cores. The heterogeneous platform has logic to facilitate sharing of pointers to a location of a memory shared by the CPU and the GPU. By sharing pointers in the heterogeneous platform, the data or information sharing between different cores in the heterogeneous platform can be simplified.

Abstract: Operating system virtual memory management for hardware transactional memory. A method may be performed in a computing environment where an application running on a first hardware thread has been in a hardware transaction, with transactional memory hardware state in cache entries correlated by memory hardware when data is read from or written to data cache entries. The data cache entries are correlated to physical addresses in a first physical page mapped from a first virtual page in a virtual memory page table. The method includes an operating system deciding to unmap the first virtual page. As a result, the operating system removes the mapping of the first virtual page to the first physical page from the virtual memory page table. As a result, the operating system performs an action to discard transactional memory hardware state for at least the first physical page. Embodiments may further suspend hardware transactions in kernel mode.

Abstract: Device, system, and method of executing a call to a routine within a transaction. In some embodiments an apparatus may include a memory having stored thereon compiled code corresponding to a transaction, wherein the transaction includes at least one call to a first routine of a pair of first and second mutually inverse routines, and wherein the compiled code includes a call to a first wrapped routine replacing the call to the first routine; and a runtime library including wrapper code, wherein the wrapper code, when executed in response to the call to the first wrapped routine, results in executing the call to the first routine within the transaction and undoing the call to the first routine responsive to abort of the transaction. Other embodiments are described and claimed.

Abstract: Hardware assisted transactional memory system with open nested transactions. Embodiments include a system whereby hardware acceleration of transactions can be accomplished by implementing open nested transaction in hardware which respect software locks such that a top level transaction can be implemented in software, and thus not be limited by hardware constraints typical when using hardware transactional memory systems.

Abstract: A method and apparatus for fine-grained filtering in a hardware accelerated software transactional memory system is herein described. A data object, which may have an arbitrary size, is associated with a filter word. The filter word is in a first default state when no access, such as a read, from the data object has occurred during a pendancy of a transaction. Upon encountering a first access, such as a first read, from the data object, access barrier operations including an ephemeral/private store operation to set the filter word to a second state are performed. Upon a subsequent/redundant access, such as a second read, the access barrier operations are elided to accelerate the subsequent access, based on the filter word being set to the second state to indicate a previous access occurred.

Abstract: A method and apparatus for dynamic optimization of strong atomicity barriers is herein described. During runtime compilation, code including non-transactional memory accesses that are to conflict with transactional memory accesses is patched to insert transactional barriers at the conflicting non-transactional memory accesses to ensure isolation and strong atomicity. However, barriers are omitted or removed from non-transactional memory accesses that do not conflict with transactional memory accesses to reduce barrier execution overhead.

Abstract: A example method disclosed herein comprises initiating a first optimistically balanced synchronization to acquire a lock of an object, the first optimistically balanced synchronization comprising a first optimistically balanced acquisition and a first optimistically balanced release to be performed on the lock by a same thread and at a same nesting level, releasing the lock after execution of program code covered by the lock if a stored state of the first optimistically balanced release indicates that the first optimistically balanced release is still valid, the stored state of the first optimistically balanced release being initialized prior to execution of the program code to indicate that the first optimistically balanced release is valid, and throwing an exception after execution of the program code covered by the lock if the stored state of the first optimistically balanced release indicates that the first optimistically balanced release is no longer valid.

Abstract: Methods, systems, and mediums are described for scheduling data parallel tasks onto multiple thread execution units of processing system. Embodiments of a lock-free queue structure and methods of operation are described to implement a method for scheduling fine-grained data-parallel tasks for execution in a computing system. The work of one of a plurality of worker threads is wait-free with respect to the other worker threads. Each node of the queue holds a reference to a task that may be concurrently performed by multiple thread execution units, but each on a different subset of data. Various embodiments relate to software-based scheduling of data-parallel tasks on a multi-threaded computing platform that does not perform such scheduling in hardware. Other embodiments are also described and claimed.

Abstract: A method and apparatus for accelerating a software transactional memory (STM) system is described herein. Annotation field are associated with lines of a transactional memory. An annotation field associated with a line of the transaction memory is initialized to a first value upon starting a transaction. In response to encountering a read operation in the transaction, then annotation field is checked. If the annotation field includes a first value, the read is serviced from the line of the transaction memory without having to search an additional write space. A second and third value in the annotation field potentially indicates whether a read operation missed the transactional memory or a tentative value is stored in a write space. Additionally, an additional bit in the annotation field, may be utilized to indicate whether previous read operations have been logged, allowing for subsequent redundant read logging to be reduced.

Abstract: A method and apparatus for a hybrid transactional memory system is herein described. A first transaction is executed utilizing a first style of a transactional memory system and a second transaction is executed in parallel utilizing a second style of a transactional memory system. For example, a main thread is executed utilizing an update-in place Software Transactional Memory (STM) system while a parallel thread, such as a helper thread, is executed utilizing a write buffering STM. As a result, a main thread may directly update memory locations, while a helper thread's transactional writes are buffered to ensure they do not invalidate transactional reads of the main thread. Therefore, parallel execution of threads is achieved, while ensuring at least one thread, such as a main thread, does not degrade below an amount of execution cycles it would take to execute the main thread serially.