摘要:
A processor subunit for a processor for processing data. The processor subunit includes registers, and at least one functional unit for executing instructions on data. One or more registers of the registers are connected to an input of the at least one functional unit, where each register connected to the input of the at least one functional unit which has an input multiplexer. One or more registers of the registers are connected to an output of the at least one functional unit, where each register connected to the output of the at least one functional unit which has an input multiplexer. At least one output bus is connected to at least one register. At least one input bus is connected to at least one register. The processor subunit may be used in a processor, which may be used in a data streaming accelerator.
摘要:
The invention relates to a method and system for the design and implementation of state machine engines. A first constraints checking step checks a state transition function created by a designer against constraints imposed by the implementation technology in order to detect all portions of the state transition function that are in conflict with the constraints. A subsequent conflict resolution step tries to determine one or more suggested ways to meet the conflicting constraints, by investigating how the original state transition function can be modified such that all constraints are met. A final presentation and selection step provides the designer textual and/or graphically results of the constraints check and suggested modifications. The modifications can be accepted interactively, or the state transition function can be changed manually. In the latter case, the modified state transition function will be processed starting again with the constraints checking step.
摘要:
The present, invention improves the hash table lookup operation by using a new processor cache architecture. A speculative processing of entries stored in the cache is combined with a delayed evaluation of cache entries. The speculative processing means that for each cache entry retrieved from main memory in a step of the hash table lookup operation it is assumed that it already contains the selected hash table entry. The delayed evaluation means that certain steps of the lookup operation are performed in parallel with others. In advantageous embodiments the invention can also be used in conjunction with a hierarchy of inclusive caches. The preferred embodiments of the invention involve a new approach for a transition rule cache of a BaRT-FSM controller.
摘要:
The invention relates to a method of optimizing a state transition function specification for a state machine engine based on a probability distribution for the state transitions. For the preferred embodiment of the invention, a B-FSM state machine engine accesses a transition rule memory using a processor cache. The invention allows improving the cache hit rate by exploiting the probability distribution. The N transition rules that comprise a hash table entry will be loaded in a burst mode from the main memory, from which the N transition rules are transferred to the processor cache. Because the comparison of the actual state and input values against each of the transition rules can immediately start after each of these rules has been received, the overall performance is improved as the transition rule that is most likely to be selected is the first to be transferred as part of the burst access.
摘要:
The invention relates to a method of optimizing a state transition function specification for a state machine engine based on a probability distribution for the state transitions. For the preferred embodiment of the invention, a B-FSM state machine engine accesses a transition rule memory using a processor cache. The invention allows improving the cache hit rate by exploiting the probability distribution. The N transition rules that comprise a hash table entry will be loaded in a burst mode from the main memory, from which the N transition rules are transferred to the processor cache. Because the comparison of the actual state and input values against each of the transition rules can immediately start after each of these rules has been received, the overall performance is improved as the transition rule that is most likely to be selected is the first to be transferred as part of the burst access.
摘要:
The invention relates to a method and system for the design and implementation of state machine engines. A first constraints checking step checks a state transition function created by a designer against constraints imposed by the implementation technology in order to detect all portions of the state transition function that are in conflict with the constraints. A subsequent conflict resolution step tries to determine one or more suggested ways to meet the conflicting constraints, by investigating how the original state transition function can be modified such that all constraints are met. A final presentation and selection step provides the designer textual and/or graphically results of the constraints check and suggested modifications. The modifications can be accepted interactively, or the state transition function can be changed manually. In the latter case, the modified state transition function will be processed starting again with the constraints checking step.
摘要:
A system and method of implementing multiple programmable finite state machines using a shared transition table is disclosed, the method including forming a plurality of finite state machine cores such that an amount of the plurality of finite state machine cores is unchangeable, forming a state transition array, and forming a routing network such that the forming the plurality of associated state transition elements is realized.
摘要:
A method and apparatus for processing data by a pipeline of a virtual multiple instance extended finite state machine (VMI EFSM). An input token is selected to enter the pipeline. The input token includes a reference to an EFSM instance, an extended command, and an operation code. The EFSM instance requires the resource to be available to generate an output token from the input token. In response to receiving an indication that the resource is unavailable, the input token is sent to a wait room or an initiative token containing the reference and the operation code is sent to a wait queue, and the output token is not generated. Without stalling and restarting the pipeline, another input token is processed in the pipeline while the resource is unavailable and while the input token is in the wait room or the initiative token is in the wait queue.
摘要:
A method and apparatus for processing data by a pipeline of a virtual multiple instance extended finite state machine (VMI EFSM). An input token is selected to enter the pipeline. The input token includes a reference to an EFSM instance, an extended command, and an operation code. The EFSM instance requires the resource to be available to generate an output token from the input token. In response to receiving an indication that the resource is unavailable, the input token is sent to a wait room or an initiative token containing the reference and the operation code is sent to a wait queue, and the output token is not generated. Without stalling and restarting the pipeline, another input token is processed in the pipeline while the resource is unavailable and while the input token is in the wait room or the initiative token is in the wait queue.
摘要:
Embodiments relate to selectively blocking branch instruction predictions. An aspect includes a computer system for performing selective branch prediction. The system includes memory and a processor, and the system is configured to perform a method. The method includes detecting a branch-prediction blocking instruction in a stream of instructions and blocking branch prediction of a predetermined number of branch instructions following the branch-prediction blocking instruction based on the detecting the branch-prediction blocking instruction.