146 results about "Reservation station" patented technology

A system and method for providing a digital real-time voltage droop detection and subsequent voltage droop reduction. A scheduler within a reservation station may store a weight value for each instruction corresponding to node capacitance switching activity for the instruction derived from pre-silicon power modeling analysis. For instructions picked with available source data, the corresponding weight values are summed together to produce a local current consumption value and this value is summed with any existing global current consumption values from corresponding schedulers of other processor cores yielding an activity event. The activity event is stored. Hashing functions within the scheduler are used to determine both a recent and an old activity average using the calculated activity event and stored older activity events. Instruction issue throttling occurs if either a difference between the old activity average and the recent activity average exceed a first threshold or the recent activity average exceeds a second threshold.

A data address prediction structure for a superscalar microprocessor is provided. The data address prediction structure predicts a data address that a group of instructions is going to access while that group of instructions is being fetched from the instruction cache. The data bytes associated with the predicted address are placed in a relatively small, fast buffer. The decode stages of instruction processing pipelines in the microprocessor access the buffer with addresses generated from the instructions, and if the associated data bytes are found in the buffer they are conveyed to the reservation station associated with the requesting decode stage. Therefore, the implicit memory read associated with an instruction is performed prior to the instruction arriving in a functional unit. The functional unit is occupied by the instruction for a fewer number of clock cycles, since it need not perform the implicit memory operation. Instead, the functional unit performs the explicit operation indicated by the instruction.

The invention provides a device and a method for dynamically scheduling instructions transmitted from an operation queue to a reservation station in a microprocessor. The method comprises the following: a step of writing instructions, which is to set and then write the operand states of the decoded instructions on the basis of data correlation between the decoded instructions to be written into the operation queue and effective instructions in the operation queue, as well as instruction execution results which have been written back and are being written; a step of updating the operand states, which is to update the operand state of each instruction not transmitted on the basis of the data correlation between each instruction not transmitted and the instructions being written back of instruction execution results; a step of judging to-be-transmitted instructions, which is to judge whether the to-be-transmitted instructions with all operands ready exist on the basis of the operand state of each instruction not transmitted; and a step of transmitting instructions, which is to transmit the judged to-be-transmitted instructions to the reservation station when the reservation station has vacancies. Pipeline efficiency can be effectively improved by transmitting the instructions with the operands ready to the reservation station on the basis of the data correlation between the instructions.

A processor device having a reservation station (RS) is concerned. In case the processor device has plural RS, the RS is associated with an arithmetic pipeline, and two RS make a pair. When one RS of the pair cannot dispatch an instruction to an associated arithmetic pipeline, the other RS dispatches the instruction to that arithmetic pipeline, or delivers its held instruction to the one RS. In case one RS is equipped, plural entries in the RS are divided into groups, and by dynamically changing this grouping according to the dispatch frequency of the instruction to the arithmetic pipelines or the held state of the instructions, the arithmetic pipelines are efficiently utilized. Incidentally, depending on the grouping of the plural entries in the RS, a configuration as if the plural RS were allocated to each arithmetic pipeline may be realized.

The invention describes allocating threads to resources using speculation metrics. A method, a reservation station and a processor for allocating resources to a plurality of threads based on the extent to which the instructions associated with each of the threads are speculative. The method comprises receiving a speculation metric for each thread at the reservation station. Each speculation metric represents the extent to which the instructions associated with a particular thread are speculative. The more speculative an instruction, the more likely the instruction has been incorrectly predicted by a branch predictor. The reservation station allocates functional unit resources (e.g. pipelines) to the threads based on the speculation metrics, and selects a number of instructions from one or more of the threads based on the allocation. The selected instructions are then issued to the functional unit resources.

A return address in response to a return instruction corresponding to a call instruction is stored in a return address stack when a branch history detects presence of the call instruction. When the branch history detects the presence of the return instruction before a branch reservation station completes executing the call instruction, the return address in response to the return instruction is not stored in the return address stack. If so, an output selection circuit predicts a correct return target using information stored in the return address stack.

A processor and method are described for scheduling operations for execution within a reservation station. For example, a method in accordance with one embodiment of the invention includes the operations of: classifying a plurality of operations based on the execution ports usable to execute those operations; allocating the plurality of operations into groups within a reservation station based on the classification, wherein each group is serviced by one or more execution ports corresponding to the classification, and wherein two or more entries within a group share a common read port and a common write port; dynamically scheduling two or more operations in a group for concurrent execution based on the ports capable of executing those operations and a relative age of the operations.

A method of selecting instructions to issue to a functional unit of an out-of-order superscalar processor single-threaded or multi-threaded. A reservation station classifies each instruction into one of a number of categories based on the type of instruction. Once classified, an instruction is stored in one of several instruction queues corresponding to the category in which it was classified. Instructions are then selected from one or more of the instruction queues (up to a maximum number of instructions for each particular queue) to issue to the functional unit based on a relative priority of the plurality of types of instructions. This allows certain types of instructions (eg. control transfer instructions, flag setting instructions and/or address generation instructions) to be prioritised over other types of instructions even if they are younger. A functional unit may contain a plurality of pipelines, and there may be several such functional units in a processor.

In an embodiment, a processor includes a multi-level dispatch circuit configured to supply operations for execution by multiple parallel execution pipelines. The multi-level dispatch circuit may include multiple dispatch buffers, each of which is coupled to multiple reservation stations. Each reservation station may be coupled to a respective execution pipeline and may be configured to schedule instruction operations (ops) for execution in the respective execution pipeline. The sets of reservation stations coupled to each dispatch buffer may be non-overlapping. Thus, if a given op is to be executed in a given execution pipeline, the op may be sent to the dispatch buffer which is coupled to the reservation station that provides ops to the given execution pipeline.

The invention discloses a NAND flash memory controller supporting operation out-of-order execution, and aims to improve memory access data throughput. The controller consists of a preprocessing module, a reservation station module, a dynamic scheduling module and a postprocessing module, wherein the preprocessing module consists of an address recognition unit, three multiplexers, a first data queue group and a first logic OR gate; the reservation station module consists of two multiplexers, a table entry number register, a validity monitoring unit, a relevance judgment unit, a substrate state setting unit, an output judgment unit, a control logic unit and a reservation station table; the dynamic scheduling module consists of five multiplexers, M state machines, an arbiter and a second logic OR gate; and the postprocessing module consists of a comparator, a second data queue group and a multiplexer. By the controller, the memory access operation of different substrates of a memory chip can be subjected to parallel execution, the working parallelism of the substrates is improved, and the memory access data throughput is effectively improved.

A system and method of processing a predicated instruction is disclosed. A consumer instruction and a predicated instruction are received in an reservation station of an out-order processor. The consumer instruction depends on a result of the predicated instruction. The predicated instruction is dispatched to an execution unit for execution. The executed predicate instruction is stored in a re-order buffer.

An apparatus including first and second reservation stations. The first reservation station dispatches a load micro instruction, and indicates on a hold bus if the load micro instruction is a specified load micro instruction directed to retrieve an operand from a prescribed resource other than on-core cache memory. The second reservation station is coupled to the hold bus, and dispatches one or more younger micro instructions therein that depend on the load micro instruction for execution after a number of clock cycles following dispatch of the first load micro instruction, and if it is indicated on the hold bus that the load micro instruction is the specified load micro instruction, the second reservation station is configured to stall dispatch of the one or more younger micro instructions until the load micro instruction has retrieved the operand. The plurality of non-core resources includes a random access memory, programmed via a Joint Test Action Group interface with the plurality of specified load instructions corresponding to the out-of-order processor which, upon initialization, accesses the random access memory to determine said plurality of specified load instructions.

A method and an apparatus for providing quality of service in a multi-processing element environment based on priority is herein described. Consumption of resources, such as a reservation station and a pipeline, are biased towards a higher priority processing element. In a reservation station, mask elements are set to provide access for higher priority processing elements to more reservation entries. In a pipeline, bias logic provides a ratio of preference for selection of a high priority processing element.

The present invention provides a method and apparatus for processing instructions in which the time allowed for the execution of an instruction is dynamically allocated. The allocation of time for execution of instruction occurs after the instruction is sent to the execution unit. The execution unit determines whether it can complete the instruction during the current processor cycle. In response to an ability to complete the instruction within the current processor cycle, the execution unit issues a busy signal to the reservation station. The reservation station continues to hold the next instruction until the execution until is capable of processing it.

Managing dataflow execution of loop instructions by out-of-order processors (OOPs), and related circuits, methods, and computer-readable media are disclosed. In one aspect, a reservation station circuit is provided. The reservation station circuit includes multiple reservation station segments, each storing a loop instruction of a loop of a computer program. Each reservation station segment also stores an instruction execution credit indicating whether the corresponding loop instruction may be provided for dataflow execution. The reservation station circuit further includes a dataflow monitor that distributes an initial instruction execution credit to each reservation station segment. As each loop iteration is executed, each reservation station segment determines whether the instruction execution credit indicates that the loop instruction for the reservation station segment may be provided for dataflow execution. If so, the reservation station segment provides the loop instruction for dataflow execution, and adjusts the instruction execution credit for the reservation station segment.

The invention discloses a task-level out-of-order multi-issue scheduler and a scheduling method thereof. The scheduler is characterized by comprising a reservation station, an option awakening unit and a calculation resource management unit, wherein the reservation station comprises a write address management unit, a storage space and a reservation station state table; the option awakening unit comprises a chronological table, a ready inquiry unit and a ready counter; and the calculation resource management unit comprises a calculation resource table, an allocation unit and a recovery unit. The throughput and resource utilization efficiency of the scheduler can be increased, so that the task instruction issue efficiency is increased, and the system performance is improved.

The present invention includes a decode section for simultaneously holding a plurality of instructions in one thread at a time and for decoding the held instructions; an execution pipeline capable of simultaneously executing each processing represented by the respective instructions belonging to different threads and decoded by the decode section; a reservation station for receiving the instructions decoded by the decode section and holding the instructions, if the decoded instructions are of sync attribute, until executable conditions are ready and thereafter dispatching the decoded instructions to the execution pipeline; a pre-decode section for confirming by a simple decoding, prior to decoding by the decode section, whether or not the instructions are of sync attribute; and an instruction buffer for suspending issuance to the decode section and holding the instructions subsequent to an instruction of sync attribute.

The invention provides a data read-write scheduler and a reservation station used for vector operation. In the data read-write scheduler, a read instruction buffer module and a write-in instruction buffer module are arranged, the read instruction buffer module and the write-in instruction buffer module mutually detect conflict instructions, when conflict exists, execution of the instruction is paused, the instruction is re-executed when the time is ripe, and therefore writing after reading conflict and reading after writing conflict between instructions are solved, the fact that correct data is provided for a vector operation part is guaranteed, and the data read-write scheduler has good promotion and application values.

An apparatus including first and second reservation stations. The first reservation station dispatches a load micro instruction, and indicates on a hold bus if the load micro instruction is a specified load micro instruction directed to retrieve an operand from a prescribed resource other than on-core cache memory, where the specified load instruction requires more than a first number of clock cycles to retrieve the operand. The second reservation station is coupled to the hold bus, and dispatches one or more younger micro instructions therein that depend on the load micro instruction for execution after a number of clock cycles following dispatch of the first load micro instruction, and if it is indicated on the hold bus that the load micro instruction is the specified load micro instruction, the second reservation station is configured to stall dispatch of the one or more younger micro instructions until the load micro instruction has retrieved the operand.