Abstract: Described herein is a neural network accelerator with a set of neural processing units and an instruction set for execution on the neural processing units. The instruction set is a compact instruction set including various compute and data move instructions for implementing a neural network. Among the compute instructions are an instruction for performing a fused operation comprising sequential computations, one of which involves matrix multiplication, and an instruction for performing an elementwise vector operation. The instructions in the instruction set are highly configurable and can handle data elements of variable size. The instructions also implement a synchronization mechanism that allows asynchronous execution of data move and compute operations across different components of the neural network accelerator as well as between multiple instances of the neural network accelerator.

Abstract: A method, apparatus, and program product facilitate the automation of an oil & gas process, e.g., a drilling process, through the use of a dynamic phase machine incorporating multiple autonomous agents.

Abstract: A state machine engine includes a state vector system. The state vector system includes an input buffer configured to receive state vector data from a restore buffer and to provide state vector data to a state machine lattice. The state vector system also includes an output buffer configured to receive state vector data from the state machine lattice and to provide state vector data to a save buffer.

Abstract: A method of monitoring thread execution within a multicore processor architecture which comprises a plurality of interconnected processor elements for processing the threads, the method comprising receiving a plurality of thread parameter indicators of one or more parameters relating to the function and/or identity and/or execution location of a thread or threads, comparing at least one of the thread parameter indicators with a first plurality of predefined criteria each representative of an indicator of interest, and generating an output consequential upon thread parameter indicators which have been identified to be of interest as a result of the said comparison.

Abstract: The present invention concerns a new category of integrated circuitry and a new methodology for adaptive or reconfigurable computing. The preferred IC embodiment includes a plurality of heterogeneous computational elements coupled to an interconnection network. The plurality of heterogeneous computational elements include corresponding computational elements having fixed and differing architectures, such as fixed architectures for different functions such as memory, addition, multiplication, complex multiplication, subtraction, configuration, reconfiguration, control, input, output, and field programmability. In response to configuration information, the interconnection network is operative in real-time to configure and reconfigure the plurality of heterogeneous computational elements for a plurality of different functional modes, including linear algorithmic operations, non-linear algorithmic operations, finite state machine operations, memory operations, and bit-level manipulations.

Abstract: The present invention concerns a new category of integrated circuitry and a new methodology for adaptive or reconfigurable computing. The preferred IC embodiment includes a plurality of heterogeneous computational elements coupled to an interconnection network. The plurality of heterogeneous computational elements include corresponding computational elements having fixed and differing architectures, such as fixed architectures for different functions such as memory, addition, multiplication, complex multiplication, subtraction, configuration, reconfiguration, control, input, output, and field programmability. In response to configuration information, the interconnection network is operative in real-time to configure and reconfigure the plurality of heterogeneous computational elements for a plurality of different functional modes, including linear algorithmic operations, non-linear algorithmic operations, finite state machine operations, memory operations, and bit-level manipulations.

Abstract: The present invention concerns a new category of integrated circuitry and a new methodology for adaptive or reconfigurable computing. The preferred IC embodiment includes a plurality of heterogeneous computational elements coupled to an interconnection network. The plurality of heterogeneous computational elements include corresponding computational elements having fixed and differing architectures, such as fixed architectures for different functions such as memory, addition, multiplication, complex multiplication, subtraction, configuration, reconfiguration, control, input, output, and field programmability. In response to configuration information, the interconnection network is operative in real-time to configure and reconfigure the plurality of heterogeneous computational elements for a plurality of different functional modes, including linear algorithmic operations, non-linear algorithmic operations, finite state machine operations, memory operations, and bit-level manipulations.

Abstract: In certain embodiments, a digital signal processor (DSP) has multiple arithmetic logic units and a register module. The DSP is adapted to generate a message digest H from a message M in accordance with the SHA-1 standard, where M includes N blocks M(i), i=1, . . . , N, and the processing of each block M(i) includes t iterations of processing words of message schedule {Wt}. In each iteration possible, the DSP uses free operations to precalculate Wt and working variable values for use in the next iteration. In addition, in each iteration possible, the DSP rotates the registers associated with particular working variables to reduce operations that merely copy unchanged values from one register to another.

Abstract: A state machine for generating signals configured for generating different signals according to the current state of the machine. The state machine is configured to change state both as a function of an internal timer and as a function of signals representative of events external to the state machine.

Abstract: In some embodiments, new clock gating approaches, referred hereafter as State Transition Gating (STG) methods and circuits are provided. In areas of circuit designs including sequential elements, the use of STG may be used to reduce dynamic power consumption.

Abstract: A method and system for executing 32-bit flat address programs during a System Management Interrupt. The system provides a 16-bit SMI routine that is given control when an SMI occurs. That routine initially saves the state of the processor and then executes an instruction to switch to protected mode. When in protected mode, the routine transfers control to 32-bit code. The 32-bit code uses a global descriptor table that is different from that used by the interrupted operating system. When the 32-bit code completes, it restores the saved processor state and returns from the interrupt by executing an RSM instruction.

Abstract: Disclosed herein is a latch circuit including a switching circuit for switching output/non-output of an externally inputted external signal based on a predetermined control signal, a state retaining circuit for inputting a signal outputted from the switching circuit as an input signal, and retaining the state of the logical level of an output signal that is outputted based on the input signal, and a clear circuit for changing the logical level of the input signal to a clear level based on a clear signal.

Abstract: Method and apparatus for indicating to a coprocessor when the coprocessor can update internal register content thereof without negative repercussion to a processor is described. A controller is coupled between the coprocessor and a processor, where the controller is configured with a state machine to track the instruction through pipeline stages of the processor.

Abstract: A state machine may have a sequence that is called by multiple threads within the state machine. Prior to calling the sequence, an address specific to the current state is stored in an address register. After the sequence has executed, the address register is queried and the thread may continue. Many different threads may call the sequence. In more complex hardware implemented state machines, the total number of gates may be reduced significantly.

Abstract: Methods and related computer program products, systems, and devices for using a NAND flash as a program ROM are disclosed.

Abstract: A method and system for executing 32-bit flat address programs during a System Management Interrupt. The system provides a 16-bit SMI routine that is given control when an SMI occurs. That routine initially saves the state of the processor and then executes an instruction to switch to protected mode. When in protected mode, the routine transfers control to 32-bit code. The 32-bit code uses a global descriptor table that is different from that used by the interrupted operating system. When the 32-bit code completes, it restores the saved processor state and returns from the interrupt by executing an RSM instruction.

Abstract: A method and apparatus to control logic sections of a pipeline instruction processor is disclosed. A state machine is provided that models the flow of instructions through the pipeline. The state machine is capable of modeling execution for all combinations of instruction types that may be present within the pipeline at a given time. The state machine also models various events that affect the way instruction execution is overlapped within the pipeline, and other system occurrences that may cause the termination of some processing activity within the pipeline. The state machine provides signals to control the various logic sections. These signals may be used to determine whether the results of processing activity within the logic sections should be retained or discarded.

Abstract: One disclosed embodiment may comprise an application specific integrated circuit (ASIC). The ASIC includes memory that stores condition data defining conditions for enabling transitions among a plurality of states and next state data defining a next state associated with each of the respective conditions. A state machine circuit employs the condition data and the next state data to transition from a current state of the state machine circuit to a next state as a function of applying at least one condition relative to input data. The at least one condition is defined by condition data that is associated with the current state. The state machine circuit associates next state data with the at least one condition based on the current state of the state machine circuit. A control circuit provides a trigger signal in response to the current state of the state machine circuit transitioning to at least one predefined state of the plurality of states.

Abstract: A processing engine, such as a digital signal processor, includes an execution mechanism, a repeat count register and a repeat count index register. The execution mechanism is operable for a repeat instruction to initialize the repeat count index register with the content of the repeat count register, and to modify the content of the repeat count register. The repeat instruction comprises two parts, the first of which initializes the repeat count index register and initiates repeat of a subsequent instruction, and the second part of which modifies the content of the repeat count register.

Abstract: At least one exemplary embodiment comprises a system comprising an arithmetic logic unit; a memory comprising a pre-computed table of target pulse widths, changes in pulse width, and pulse counts distributed according to a constrained semi-logarithmic distribution, the memory connected to the arithmetic logic unit via a pipeline mechanism; and a state machine adapted to load each of the target pulse widths and changes in pulse width from the memory into the arithmetic logic unit at pre-determined intervals of pulse count while maintaining control of a pulse width generated by the arithmetic logic unit.

Abstract: An invention is disclosed for a layer structure that facilitates configuring a Fiber Channel driver. In one embodiment, the layer structure includes a hardware layer directory that includes code for communicating with a Fiber Channel controller. In addition, a wrapper layer directory is included in the layer structure. The wrapper layer directory includes code for communicating with the code associated with the hardware layer directory, and also includes a wrapper header file that defines a particular value setting in a first state, such as a compiler directive set a particular value. The layer structure further includes a global header directory that defines a group of value settings. The group of value settings is defined for the code associated with each of the hardware directory and the wrapper layer directory. The particular value setting in the first state is also included in the group of value settings.

Abstract: A programmable controller for controlling a multi-channel sequential processing of states based on a prescribed state transitions. The microcontroller has a state register for each of the channels for holding the state data to be processed in the channel in the next processing period, and upon receipt of a sampling clock and a channel processing request, executes a program to process the states in each of the channels based on the states data held in the associated state register. The microcontroller may execute a program having a complex sequence in a reduced number of steps.

Abstract: A mispredicted path side memory is configured to be coupled to a stage in an instruction pipeline. As instructions advance through the pipeline, a result from the stage is stored into the mispredicted path side memory. The result is restored from the mispredicted path side memory into a pipeline stage when a branch is mispredicted.

Abstract: A processor is provided with a set of instructions formed, in general, of an operation section and an operand section. For a special control instruction, the operand section is transmitted to the operation blocks along a bypass path separate from the normal path in which normal instructions are interpreted. In this way, an extension of the set of instructions can be achieved for tailoring the set of instructions to the user's own requirements. Consequently, the processor control unit should be capable of coupling its outputs to its inputs upon receiving one such instruction, thereby to transfer such internal operation control signals without interpretation.

Abstract: A finite state machine that implements the ETSI Intelligent Network Capability Set 1 for INAP protocol using a first finite state machine connected to a second finite state machine. To accomplish this implementation, the first finite state machine receives inputs and makes initial state transition choices. Once the first finite state machine has made an initial choice, the second finite state machine checks to see if the first finite state machine has transitioned to the appropriate state and, if it has not, give feedback to the finite state machine to the appropriate state. Once the appropriate state is reached, the outputs of the first finite state machine are valid and available. The finite state machines may be realized as physical logical devices or virtually using VFSM technology.

Abstract: Two different types of memory are integrated on the same type of integrated circuit. A microcontroller is associated with each of these memories. In order that the independence of operation of these microcontrollers may be ensured, they are each provided with time delay circuits whose role is to maintain the read or write operation of one microcontroller when another is selected.

Abstract: A packed data structure processor (25) is disclosed. The packed data structure processor (25) includes a register file (24) of multiple registers (REG0 through REG31), each of which is connected to an input of each of a plurality of operand multiplexers (26). Each operand multiplexer (26) is associated with a shift/mask circuit (28), which permits the selection of a particular portion (e.g., BYIE, WORD, DWORD) of the contents of a selected register file, for use as an operand. An arithmetic logic unit (ALU) (30) performs data processing operations upon the operands, and presents results on writeback bus (WBBUS), to external memory (18) over a memory interface (37), or to a register file (42) associated with other circuitry (44) over a coprocessor interface (41). A destination selector (40) is capable of writing to only a selected portion of a selected register, thus permitting a packed data structure to be present within the register file (24).

Abstract: A system and method for thermal overload detection and protection for a processor which allows the processor to run at near maximum potential for the vast majority of its execution life. This is effectuated by the provision of circuitry to detect when the processor has exceeded its thermal thresholds and which then causes the processor to automatically reduce the clock rate to a fraction of the nominal clock while execution continues. When the thermal condition has stabilized, the clock may be raised in a stepwise fashion back to the nominal clock rate. Throughout the period of cycling the clock frequency from nominal to minimum and back, the program continues to be executed. Also provided is a queue activity rise time detector and method to control the rate of acceleration of a functional unit from idle to full throttle by a localized stall mechanism at the boundary of each stage in the pipe.

Abstract: A microprocessor includes a decoder, a queue, and a renamer. The decoder is adapted to receive a program instruction and decode the program instruction to provide a first decoded instruction. The first decoded instruction includes a plurality of instruction bits. The queue is coupled to the decoder and adapted to store the first decoded instruction. The renamer has a first input port and a first and second output port. The renamer is coupled to the queue and adapted to receive the first decoded instruction at the input port, provide the first decoded instruction on the first output port, change at least one of the instruction bits to generate a second decoded instruction, and provide the second decoded instruction on the second output port. A method for expanding program instructions in a microprocessor having a renamer is provided. The renamer includes a first input port and first and second output ports. The method includes receiving a first decoded instruction in the first input port.

Abstract: Two different types of memory are integrated on the same type of integrated circuit. A microcontroller is associated with each of these memories. In order that the independence of operation of these microcontrollers may be ensured, they are each provided with time delay circuits whose role is to maintain the read or write operation of one microcontroller when another is selected.

Abstract: The invention relates to a processor having a data flow unit for processing data in a plurality of steps. In one version, the data flow unit includes a plurality of consecutive stages which include logic for performing steps of the data processing, the stages being coupled together by a data path, at least one stage being coupled to a transceiver which causes data to be provided to the stage for processing or to bypass the stage unprocessed in response to a stage enable signal; a synchronizer which receives processed data from the stages and causes the processed data to be provided to external logic in synchronization with a clock signal.

Abstract: A system and method for thermal overload detection and protection for a processor which allows the processor to run at near maximum potential for the vast majority of its execution life. This is effectuated by the provision of circuitry to detect when the processor has exceeded its thermal thresholds and which then causes the processor to automatically reduce the clock rate to a fraction of the nominal clock while execution continues. When the thermal condition has stabilized, the clock may be raised in a stepwise fashion back to the nominal clock rate. Throughout the period of cycling the clock frequency from nominal to minimum and back, the program continues to be executed. Also provided is a queue activity rise time detector and method to control the rate of acceleration of a functional unit from idle to full throttle by a localized stall mechanism at the boundary of each stage in the pipe.