Abstract: A resource-lock monitor detects when processors in a multi-processor system are stuck waiting for access to a shared resource. A lock-monitor register has a lock bit and a sticky-lock bit for each processor being monitored. The lock and the sticky-lock bits are both set when the processor executes a lock instruction that also sends a lock-request to a resource arbiter. The lock bit is cleared when the resource arbiter grants access to the processor, but the sticky-lock bit remains set until sticky-lock bits are cleared by monitoring software at the end of a monitoring period. At the end of each monitoring period, monitoring software reads the lock and sticky-lock bits and finds a locked processor when a processor's lock bit is still set, but its sticky-lock bit is cleared. When the locked processor remains locked at the end of another monitoring period, an error handler resets the locked processor.

Abstract: Load and store operations in computer systems are extended to provide for Stream Load and Store and Masked Load and Store. In Stream operations, a CPU executes a Stream instruction that indicates, by appropriate arguments, a first address in memory or a first register in a register file from whence to begin reading data entities, and a first address or register from whence to begin storing the entities, and a number of entities to be read and written. In Masked Load and Masked Store operations stored masks are used to indicate patterns relative to first addresses and registers for loading and storing. Bit-string vector methods are taught for masks.

Abstract: In a data-packet processor, a configurable queueing system for packet accounting during processing has a plurality of queues arranged in one or more clusters, an identification mechanism for creating a packet identifier for arriving packets, insertion logic for inserting packet identifiers into queues and for determining into which queue to insert a packet identifier, and selection logic for selecting packet identifiers from queues to initiate processing of identified packets, downloading of completed packets, or for requeueing of the selected packet identifiers.

Abstract: A background memory manager (BMM) for managing a memory in a data processing system has circuitry for transferring data to and from an outside device and to and from a memory, a memory state map associated with the memory, and a communication link to a processor. The BMM manages the memory, determining if each data structure fits into the memory, deciding exactly where to place the data structure in memory, performing all data transfers between the outside device and the memory, and maintaining the memory state map according to memory transactions made, and informing the processor of new data and its location. In preferred embodiments the BMM, in the process of storing data structures into the memory, provides an identifier for each structure to the processor. The system is particularly applicable to Internet packet processing in packet routers.

Abstract: A dynamic multistreaming processor has instruction queues, each instruction queue corresponding to an instruction stream, and execution units. The dynamic multistreaming processor also has a dispatch stage to select at least one instruction from one of the instruction queues and to dispatch the selected at least one instruction to one of the execution units. Lastly the dynamic multistreaming processor has a queue counter, associated with each instruction queue, for indicating the number of instructions in each queue, and a fetch counter, associated with each instruction queue, for indicating an address from which to obtain instructions when the associated instruction queue is not full. The dynamic multistreaming processor might also have fetch counters for indicating a next instruction address from which to obtain at least one instruction when the associated instruction queue is not full. The dynamic multistreaming processor could also have a second counter for indicating a next instruction address.

Abstract: A pipelined multistreaming processor has an instruction source, a plurality of streams fetching instructions from the instruction source, a dispatch stage for selecting and dispatching instructions to a set of execution units, a set of instruction queues having one queue associated with each stream in the plurality of streams, and located in the pipeline between the instruction cache and the dispatch stage, and a select system for selecting streams in each cycle to fetch instructions from the instruction cache. The processor is characterized in that the select system selects one or more streams in each cycle for which to fetch instructions from the instruction cache, and in that the number of streams selected for which to fetch instructions in each cycle is fewer than the number of streams in the plurality of streams.

Abstract: Load and store operations in computer systems are extended to provide for Stream Load and Store and Masked Load and Store. In Stream operations a CPU executes a Stream instruction that indicates by appropriate arguments a first address in memory or a first register in a register file from whence to begin reading data entities, and a first address or register from whence to begin storing the entities, and a number of entities to be read and written. In Masked Load and Masked Store operations stored masks are used to indicate patterns relative to first addresses and registers for loading and storing. Bit-string vector methods are taught for masks.

Abstract: A multi-streaming processor has a plurality of streams for streaming one or more instruction threads, a set of functional resources for processing instructions from streams, and a lock mechanism for locking selected memory locations shared by streams of the processor, the hardware-lock mechanism operating to set a lock when an atomic memory sequence is started and to clear a lock when an atomic memory sequence is completed. In preferred embodiments the lock mechanism comprises one or more storage locations associated with each stream of the processor, each storage location enabled to store a memory address a lock bit, and a stall bit. Methods for practicing the invention using the apparatus are also taught.

Abstract: A pipelined multistreaming processor has an instruction source, a plurality of streams fetching instructions from the instruction source, a dispatch stage for selecting and dispatching instructions to a set of execution units, a set of instruction queues having one queue associated with each stream in the plurality of streams, and located in the pipeline between the instruction cache and the dispatch stage, and a select system for selecting streams in each cycle to fetch instructions from the instruction cache. The processor is characterized in that the select system selects one or more streams in each cycle for which to fetch instructions from the instruction cache, and in that the number of streams selected for which to fetch instructions in each cycle is fewer than the number of streams in the plurality of streams.

Abstract: A logic system in a data packet processor is provided for selecting and releasing one of a plurality of contexts. The selected and released context is dedicated for enabling the processing of interrupt service routines corresponding to interrupts generated in data packet processing and pending for service. The system comprises, a first determination logic for determining control status of all of the contexts, a second determination logic for determining if a context is idle or not, a selection logic for selecting a context and a context release mechanism for releasing the selected context. Determination by the logic system that all contexts are singularly owned by an entity not responsible for packet processing and that at least one of the contexts is idle, triggers immediate selection and release of an idle one of the at least one idle contexts to an entity responsible for packet processing.

Abstract: A system is provided for enabling a non-speculative pre-fetch operation for processing instructions to be performed in the background ahead of immediate packet processing by a packet processor. The system comprises a packet-management unit for accepting data packets and enqueuing them for processing, a processor unit for processing the data packets, a processor core memory for holding context registers and functional units for processing, a memory for holding a plurality of instruction threads and a software-configurable hardware table for relating queues to pointers to beginnings of instruction threads. The packet-management unit selects an available context in the processor core for processing of a data packet, consults the table, and communicates the pointer to the processor, enabling the processor to perform the non-speculative pre-fetch for instructions.

Abstract: A multi-streaming processor has a plurality of streams for streaming one or more instruction threads, a set of functional resources for processing instructions from streams, and a lock mechanism for locking selected memory locations shared by streams of the processor, the hardware-lock mechanism operating to set a lock when an atomic memory sequence is started and to clear a lock when an atomic memory sequence is completed. In preferred embodiments the lock mechanism comprises one or more storage locations associated with each stream of the processor, each storage location enabled to store a memory address a lock bit, and a stall bit. Methods for practicing the invention using the apparatus are also taught.

Abstract: In a multi-streaming processor having a memory cache, a system for fetching instructions from individual ones of multiple streams to an instruction pipeline is provided, comprising a fetch algorithm for selecting from which stream to fetch an instruction, and a hit/miss predictor for forecasting whether a load instruction will hit or miss the cache. The prediction by the hit-miss predictor is used by the fetch algorithm in determining from which stream to fetch. A hit prediction results in a next instruction being fetched from the same stream as the instruction tested by the hit/miss predictor, while a miss prediction results in the next instruction being fetched from a different stream, if any. The predictor is also used to determine which instructions to dispatch to functional units.

Abstract: In a multi-streaming processor, a system for fetching instructions from individual ones of multiple streams to an instruction pipeline is provided, comprising a fetch algorithm for selecting from which stream to fetch an instruction, and one or more predictors for forecasting whether a load instruction will hit or miss the cache or a branch will be taken. The prediction or predictions are used by the fetch algorithm in determining from which stream to fetch. In some cases probabilities are determined and also used in decisions, and predictors may be used at either or both of fetch and dispatch stages.

Abstract: A system for managing packets incoming to a data router has a local packet memory (LPM) mapped into pre-configured memory units, to store packets for processing, an external packet memory (EPM), a first storage system to store packets in the LPM, and a second storage system to store packets in the EPM. The system is characterized in that the first storage system attempts to store all incoming packets in the LPM, and for those packets that are not compatible with the LPM, relinquishes control to the second system, which stores the LPM-incompatible packets in the EPM.

Abstract: A software program extension for a dynamic multi-streaming processor is disclosed. The extension comprising an instruction set enabling coordinated interaction between a packet management component and a core processing component of the processor. The software program comprises, a portion thereof for managing packet uploads and downloads into and out of memory, a portion thereof for managing specific memory allocations and de-allocations associated with enqueueing and dequeuing data packets, a portion thereof for managing the use of multiple contexts dedicated to the processing of a single data packet; and a portion thereof for managing selection and utilization of arithmetic and other context memory functions associated with data packet processing. The extension complements standard data packet processing program architecture for specific use for processors having a packet management unit that functions independently from a streaming processor unit.

Abstract: A hardware/software system is provided for allocating memory in the form of a buffer zone surrounding a data packet to be stored in the memory. The hardware/software system comprises, first and second registers for storing separate values representing in one register, an amount of memory preceding the first line of the data packet to be stored and in the other the amount succeeding the last line of the packet to be stored, a hardware mechanism for allocating the memory according to computational results computed using the register values and the size of a data packet to be stored, and software for processing stored data packet and for writing any new growth data into the designated buffer zones surrounding the data packet.

Abstract: A multi-streaming processor has a plurality of streams for streaming one or more instruction threads, a set of functional resources for processing instructions from streams, and interrupt handler logic. The logic detects and maps interrupts and exceptions to one or more specific streams. In some embodiments one interrupt or exception may be mapped to two or more streams, and in others two or more interrupts or exceptions may be mapped to one stream. Mapping may be static and determined at processor design, programmable, with data stored and amendable, or conditional and dynamic, the interrupt logic executing an algorithm sensitive to variables to determine the mapping. Interrupts may be external interrupts generated by devices external to the processor software (internal) interrupts generated by active streams, or conditional, based on variables. After interrupts are acknowledged streams to which interrupts or exceptions are mapped are vectored to appropriate service routines.

Abstract: A logic system in a data packet processor is provided for selecting and releasing one of a plurality of contexts. The selected and released context is dedicated for enabling the processing of interrupt service routines corresponding to interrupts generated in data packet processing and pending for service. The system comprises, a first determination logic for determining control status of all of the contexts, a second determination logic for determining if a context is idle or not, a selection logic for selecting a context and a context release mechanism for releasing the selected context. Determination by the logic system that all contexts are singularly owned by an entity not responsible for packet processing and that at least one of the contexts is idle, triggers immediate selection and release of an idle one of the at least one idle contexts to an entity responsible for packet processing.

Abstract: A system for managing packets incoming to a data router has a local packet memory (LPM) mapped into pre-configured memory units, to store packets for processing, an external packet memory (EPM), a first storage system to store packets in the LPM, and a second storage system to store packets in the EPM. The system is characterized in that the first storage system attempts to store all incoming packets in the LPM, and for those packets that are not compatible with the LPM, relinquishes control to the second system, which stores the LPM-incompatible packets in the EPM.