Abstract: A system and method for computing in memory with artificial neurons. According to an embodiment of the present disclosure, there is provided a system, including: a computer-readable memory; a neuron processing element communicatively connected to the computer-readable memory, the neuron processing element including: a plurality of configurable processing circuits each having a plurality of outputs and a plurality of inputs; and a network connecting one or more of the outputs of the configurable processing circuits to one or more of the inputs of the configurable processing circuits, each of the configurable processing circuits including: an artificial neuron having a plurality of inputs; and a register connected to the inputs of the artificial neuron.

Abstract: Disclosed is neural network circuitry having a first plurality of logic cells that is interconnected to form neural network computation units that are configured to perform approximate computations. The neural network circuitry further includes a second plurality of logic cells that is interconnected to form a controller hierarchy that is interfaced with the neural network computation units to control pipelining of the approximate computations performed by the neural network computational units. In some embodiments the neural network computation units include approximate multipliers that are configured to perform approximate multiplications that comprise the approximate computations. The approximate multipliers include preprocessing units that reduce latency while maintaining accuracy.

Abstract: Threshold logic gates using flash transistors are provided. In an exemplary aspect, flash threshold logic (FTL) provides a novel circuit topology for realizing complex threshold functions. FTL cells use floating gate (flash) transistors to realize all threshold functions of a given number of variables. The use of flash transistors in the FTL cell allows a fine-grained selection of weights, which is not possible in traditional complementary metal-oxide-semiconductor (CMOS)-based threshold logic cells. Further examples include a novel approach for programming the weights of an FTL cell for a specified threshold function using a modified perceptron learning algorithm.

Abstract: A field-programmable gate array (FPGA) with reconfigurable threshold logic gates for improved performance, power, and area (PPA) is provided. This disclosure describes a new architecture for an FPGA, referred to as threshold logic FPGA (TLFPGA), that integrates a conventional lookup table (LUT) with a complementary metal-oxide-semiconductor (CMOS) digital implementation of a binary perceptron, referred to as a threshold logic cell (TLC). The TLFPGA design described herein, combined with a new logic mapping algorithm that exploits the presence of both conventional LUTs and TLCs within the basic logic element (BLE) block, achieves significant improvements in all the metrics of PPA. The TLCs of embodiments described herein are capable of implementing a complex threshold function, which if implemented using conventional gates would require several levels of logic gates. The TLCs only require seven static random-access memory (SRAM) cells and are significantly faster than the conventional LUTs.

Abstract: A configurable binary neural network (BNN) application-specific integrated circuit (ASIC) using a network of programmable threshold logic standard cells is provided. A new architecture is presented for a BNN that uses an optimal schedule for executing the operations of an arbitrary BNN. This architecture, also referred to herein as TULIP, is designed with the goal of maximizing energy efficiency per classification. At the top-level, TULIP consists of a collection of unique processing elements (TULIP-PEs) that are organized in a single instruction, multiple data (SIMD) fashion. Each TULIP-PE consists of a small network of binary neurons, and a small amount of local memory per neuron. Novel algorithms are presented herein for mapping arbitrary nodes of a BNN onto the TULIP-PEs. Comparison results show that TULIP is consistently 3× more energy-efficient than conventional designs, without any penalty in performance, area, or accuracy.

Abstract: A field-programmable gate array (FPGA) with reconfigurable threshold logic gates for improved performance, power, and area (PPA) is provided. This disclosure describes a new architecture for an FPGA, referred to as threshold logic FPGA (TLFPGA), that integrates a conventional lookup table (LUT) with a complementary metal-oxide-semiconductor (CMOS) digital implementation of a binary perceptron, referred to as a threshold logic cell (TLC). The TLFPGA design described herein, combined with a new logic mapping algorithm that exploits the presence of both conventional LUTs and TLCs within the basic logic element (BLE) block, achieves significant improvements in all the metrics of PPA. The TLCs of embodiments described herein are capable of implementing a complex threshold function, which if implemented using conventional gates would require several levels of logic gates. The TLCs only require seven static random-access memory (SRAM) cells and are significantly faster than the conventional LUTs.

Abstract: A non-volatile logic device for energy-efficient logic state restoration is disclosed. The non-volatile logic device incorporates a volatile flip-flop and a non-volatile storage unit to achieve on-chip non-volatile storage. The non-volatile logic device further allows for a backup time to be determined on a per-chip basis, resulting in minimizing energy wastage and satisfying a given yield constraint.

Abstract: Disclosed is neural network circuitry having a first plurality of logic cells that is interconnected to form neural network computation units that are configured to perform approximate computations. The neural network circuitry further includes a second plurality of logic cells that is interconnected to form a controller hierarchy that is interfaced with the neural network computation units to control pipelining of the approximate computations performed by the neural network computational units. In some embodiments the neural network computation units include approximate multipliers that are configured to perform approximate multiplications that comprise the approximate computations. The approximate multipliers include preprocessing units that reduce latency while maintaining accuracy.

Abstract: This disclosure relates generally to digital synchronous circuits that introduce clock skew without requiring clock buffers in a clock network. In one embodiment, the digital synchronous circuit includes a first flip flop and a second flip flop. The first flip flop is synchronized to be transparent and to be opaque in accordance with a first clock signal while the second flip flop is configured such that the second flip flop is synchronized to be transparent and to be opaque in accordance with a second clock signal. However, the second flip flop is configured to generate the first clock signal such that the second flip flop provides the first clock signal in a first clock state in response the second flip flop becoming transparent and provides the first clock signal in a second clock state in response the second flip flop becoming opaque thereby providing a clock skew without clock buffers.

Abstract: A sequential state element (SSE) is disclosed. In one embodiment, an SSE includes a differential sense flip flop (DSFF) and a completion detection circuit (CDC) operably associated with the DSFF. The DSFF is configured to generate a differential logical output. During a normal operational mode, the DSFF is synchronized by a clock signal to provide a differential logical output in a differential output state in accordance with a data input or in a precharge state based on the clock signal. The differential logical output is provided in a differential output state in accordance with a test input during a scan mode. The CDC is configured to generate a test enable input during the scan mode that indicates the scan mode once the differential logical output is in the differential output state. Accordingly, another SSE can be asynchronously triggered to operate in the scan mode without a separate scan clock.

Abstract: A non-volatile logic device for energy-efficient logic state restoration is disclosed. The non-volatile logic device incorporates a volatile flip-flop and a non-volatile storage unit to achieve on-chip non-volatile storage. The non-volatile logic device further allows for a backup time to be determined on a per-chip basis, resulting in minimizing energy wastage and satisfying a given yield constraint.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for obtaining an electrocardiographic (ECG) signal of a user; obtaining a feature vector of the ECG signal of the user with neural network based feature extraction. Comparing the feature vector of the ECG signal with a stored feature vector of a registered user. Authenticating the user in response to determining that a similarity of the ECG feature vector of the ECG signal and the stored ECG feature vector of the registered user exceeds a pre-defined threshold value.

Abstract: A differential sense flip flop (DSFF) that is named the Kulkarni Vrudhula flip flop (KVFF) is disclosed. In one embodiment, the DSFF includes a differential sense amplifier and an SR latch. The differential sense amplifier includes a first amplifier branch having a first output node, a second amplifier branch having a second, a first switchable discharge path, and a second switchable discharge path. The first switchable discharge path is closed to discharge the first output node when first output node is being discharged by the first amplifier branch and the second switchable discharge path is closed to discharge the second output node when second output node is being discharged by the second amplifier branch. This prevents the output nodes from floating and increases the reliability and robustness of the DSFF.

Abstract: A control system for use with a processor includes: (i) a controller configured to receive prediction information for a predicted temperature associated with the processor, and to determine a speed of operation for the processor based at least on a thermal model of the processor and the predicted temperature, where the speed supports an operational objective of the processor; and (ii) an error estimator that is separate from the controller, and that is configured to receive temperature information obtained from the processor operating at the speed, to determine updated prediction information based, at least in part, on the temperature information, and to output the updated prediction information to the controller.

Abstract: A differential sense flip flop (DSFF) that is named the Kulkarni Vrudhula flip flop (KVFF) is disclosed. In one embodiment, the DSFF includes a differential sense amplifier and an SR latch. The differential sense amplifier includes a first amplifier branch having a first output node, a second amplifier branch having a second, a first switchable discharge path, and a second switchable discharge path. The first switchable discharge path is closed to discharge the first output node when first output node is being discharged by the first amplifier branch and the second switchable discharge path is closed to discharge the second output node when second output node is being discharged by the second amplifier branch. This prevents the output nodes from floating and increases the reliability and robustness of the DSFF.

Abstract: A sequential state element (SSE) is disclosed. In one embodiment, an SSE includes a differential sense flip flop (DSFF) and a completion detection circuit (CDC) operably associated with the DSFF. The DSFF is configured to generate a differential logical output. During a normal operational mode, the DSFF is synchronized by a clock signal to provide a differential logical output in a differential output state in accordance with a data input or in a precharge state based on the clock signal. The differential logical output is provided in a differential output state in accordance with a test input during a scan mode. The CDC is configured to generate a test enable input during the scan mode that indicates the scan mode once the differential logical output is in the differential output state. Accordingly, another SSE can be asynchronously triggered to operate in the scan mode without a separate scan clock.

Abstract: Neuromorphic computational circuitry is disclosed that includes a cross point resistive network and line control circuitry. The cross point resistive network includes variable resistive units. One set of the variable resistive units is configured to generate a correction line current on a conductive line while other sets of the variable resistive units generate resultant line currents on other conductive lines. The line control circuitry is configured to receive the line currents from the conductive lines and generate digital vector values. Each of the digital vector values is provided in accordance with a difference between the current level of a corresponding resultant line current and a current level of the correction line current. In this manner, the digital vector values are corrected by the current level of the correction line current in order to reduce errors resulting from finite on to off conductance state ratios.

Abstract: An example process for controlling a processor may include: (i) obtaining parameters associated with operation of a processor, where each of the parameters has a different time scale; (ii) performing an iterative process to identify ones of the parameters that achieve a particular energy efficiency in the processor, where the energy efficiency of the processor corresponds to a quasi-concave function having a maximum that corresponds to the ones of the parameters; and (iii) controlling the processor using the ones of the parameters.

Abstract: A threshold logic element (TLE) is disclosed. The TLE includes a first input gate network, a second input gate network, and a differential sense amplifier. The first input gate network is configured to receive a first set of logical signals and the second input gate network configured to receive a second set of logical signals. The differential sense amplifier is operably associated with the first input gate network and the second input gate network such that the differential sense amplifier is configured to generate a differential logical output in accordance with a threshold logic function. To obfuscate the TLE, any number of obfuscated transmission gates can be provided in one or both of the input gate networks. The obfuscated transmission gates are obfuscated such that obfuscated transmission gates are incapable of effecting the threshold logic function of the TLE and thus hide the functionality of the TLE.

Abstract: This disclosure relates generally to digital synchronous circuits that introduce clock skew without requiring clock buffers in a clock network. In one embodiment, the digital synchronous circuit includes a first flip flop and a second flip flop. The first flip flop is synchronized to be transparent and to be opaque in accordance with a first clock signal while the second flip flop is configured such that the second flip flop is synchronized to be transparent and to be opaque in accordance with a second clock signal. However, the second flip flop is configured to generate the first clock signal such that the second flip flop provides the first clock signal in a first clock state in response the second flip flop becoming transparent and provides the first clock signal in a second clock state in response the second flip flop becoming opaque thereby providing a clock skew without clock buffers.