Abstract: There is provided a system and methods for segmenting datapath resources such as reorder buffers, physical registers, instruction queues and load-store queues, etc. in a microprocessor so that their size may be dynamically expanded and contracted. This is accomplished by allocating and deallocating individual resource units to each resource based on sampled estimates of the instantaneous resource needs of the program running on the microprocessor. By keeping unused datapath resources to a minimum, power and energy savings are achieved by shutting off resource units that are not needed for sustaining the performance requirements of the running program. Leakage energy and switching energy and power are reduced using the described methods.

Abstract: A file system that permits predictable accesses to file data stored on devices that may have a variable access latency dependent on the physical location of the file on the physical storage device. A variety of features that guarantee timely, real-time response to I/O file system requests that specify deadlines or other alternative required quality-of-service parameters. The file system addresses needs to accommodate the file systems of storage devices such as disks that have an access time dependant on the physical location of the data within the storage device. A two-phase, deadline-driven scheduler considers the impact of disk seek time on overall response times. Non real-time file operations may be preempted. Files may be preallocated to help avoid access delay caused by non-contiguity. Disk buffers may also be preallocated to improve real-time file system performance.

Abstract: Method and hardware apparatus are disclosed for reducing the rollback penalty on exceptions in a microprocessor executing traces of scheduled instructions. Speculative state is committed to the architectural state of the microprocessor at a series of commit points within a trace, rather than committing the state as a single atomic operation at the end of the trace.

Abstract: Reliable byte stream transfer protocols play an important role in modern networks. The present invention implements such a protocol using credits for flow control and negative acknowledgements for reliable delivery. The credit mechanism uses credit transfer in installments and is immune to the losses or corruptions of intermediate credit installments. Negative acknowledgements are used to solicit the retransmission of data stream component. The present invention provides full compatibility at the programming interface with reliable byte transfer protocols, such as TCP, and also allows fall interoperability among hosts running the aforesaid standard protocol or the protocol of the present invention.

Abstract: In distributed memory multiprocessors, communication between processing elements (PEs) can have a significant impact on the overall computation time. In addition, contention for the communication links can often make PEs wait even longer for a message than would normally be required. Because of this, it is important to minimize the effects of inter-processor communication time. The present invention reduces the execution time of a parallel program by merging messages (also called message combining, or message consolidation) after the program has already been partitioned and scheduled onto the PEs. The data from two (or more) messages are combined and sent in a single communication by locating sections of the system where a merge will affect the overall execution time, and determining before the merge takes place whether it will positively or negatively affect the system.