Embodiments are directed to systems and methods of implementing an analog neural network using a pipelined SRAM architecture (“PISA”) circuitry disposed in on-chip processor memory circuitry. The on-chip processor memory circuitry may include processor last level cache (LLC) circuitry. One or more physical parameters, such as a stored charge or voltage, may be used to permit the generation of an in-memory analog output using a SRAM array. The generation of an in-memory analog output using only word-line and bit-line capabilities beneficially increases the computational density of the PISA circuit without increasing power requirements. Thus, the systems and methods described herein beneficially leverage the existing capabilities of on-chip SRAM processor memory circuitry to perform a relatively large number of analog vector/tensor calculations associated with execution of a neural network, such as a recurrent neural network, without burdening the processor circuitry and without significant impact to the processor power requirements.

Embodiments are directed to systems and methods of implementing an analog neural network using a pipelined SRAM architecture (“PISA”) circuitry disposed in on-chip processor memory circuitry. The on-chip processor memory circuitry may include processor last level cache (LLC) circuitry. One or more physical parameters, such as a stored charge or voltage, may be used to permit the generation of an in-memory analog output using a SRAM array. The generation of an in-memory analog output using only word-line and bit-line capabilities beneficially increases the computational density of the PISA circuit without increasing power requirements. Thus, the systems and methods described herein beneficially leverage the existing capabilities of on-chip SRAM processor memory circuitry to perform a relatively large number of analog vector/tensor calculations associated with execution of a neural network, such as a recurrent neural network, without burdening the processor circuitry and without significant impact to the processor power requirements.

This application relates to computing circuitry, and in particular to analogue computing circuitry suitable for neuromorphic computing. An analogue computation unit for processing data is supplied with a first voltage from a voltage regulator which is operable in a sequence of phases to cyclically regulate the first voltage. A controller is configured to control operation of the voltage regulator and/or the analogue computation unit, such that the analogue computation unit processes data during a plurality of compute periods that avoid times at which the voltage regulator undergoes a phase transition which is one of a predefined set of phase transitions between defined phases in said sequence of phases. This avoids performing computation operations during a phase transition of the voltage regulator that could result in a transient or disturbance in the first voltage, which could adversely affect the computing.

This application relates to computing circuitry, and in particular to analogue computing circuitry suitable for neuromorphic computing. An analogue computation unit for processing data is supplied with a first voltage from a voltage regulator which is operable in a sequence of phases to cyclically regulate the first voltage. A controller is configured to control operation of the voltage regulator and/or the analogue computation unit, such that the analogue computation unit processes data during a plurality of compute periods that avoid times at which the voltage regulator undergoes a phase transition which is one of a predefined set of phase transitions between defined phases in said sequence of phases. This avoids performing computation operations during a phase transition of the voltage regulator that could result in a transient or disturbance in the first voltage, which could adversely affect the computing.

A method can be used to test an electronic circuit. The method includes applying a test stimulus signal to the input node, collecting a sequence of N-bit digital test data at the output port. The N-bit digital test data is determined by the test stimulus signal applied to the input node. The method also includes applying N-bit to R-bit lossless compression to the N-bit digital test data to obtain R-bit compressed test data (R is less than N) and making the R-bit compressed test data available in parallel format over R output pins of the circuit.

In an example, a circuit includes a first comparator, a second comparator, a pulse counter, a processor, a first ADC, and a second ADC. The first comparator has a first input coupled to a first node, a second input, and an output. The second comparator has a first input coupled to a second node, a second input, and an output. A first DAC is coupled to the second input of the first comparator. A second DAC is coupled to the second input of the second comparator. The pulse counter has a first input coupled to the output of the first comparator and a second input coupled to the output of the second comparator. The first ADC has an input coupled to the first node and an output coupled to the processor. The second ADC has an input coupled to the second node and an output coupled to the processor.

In an example, a circuit includes a first comparator, a second comparator, a pulse counter, a processor, a first ADC, and a second ADC. The first comparator has a first input coupled to a first node, a second input, and an output. The second comparator has a first input coupled to a second node, a second input, and an output. A first DAC is coupled to the second input of the first comparator. A second DAC is coupled to the second input of the second comparator. The pulse counter has a first input coupled to the output of the first comparator and a second input coupled to the output of the second comparator. The first ADC has an input coupled to the first node and an output coupled to the processor. The second ADC has an input coupled to the second node and an output coupled to the processor.

A power converter controller includes an analog to digital converter (ADC) to generate a digital representation of a feedback signal of a power converter, the feedback signal being received from a compensator of the power converter and being based on an output voltage of the power converter. A nonlinear gain block of the power converter controller receives the digital representation of the feedback signal and generates a transformed digital representation of the feedback signal using a nonlinear function. A switch control block of the power converter controller controls an on-time of a primary-side switch of the power converter based on the transformed digital representation of the feedback signal.

A power converter controller includes an analog to digital converter (ADC) to generate a digital representation of a feedback signal of a power converter, the feedback signal being received from a compensator of the power converter and being based on an output voltage of the power converter. A nonlinear gain block of the power converter controller receives the digital representation of the feedback signal and generates a transformed digital representation of the feedback signal using a nonlinear function. A switch control block of the power converter controller controls an on-time of a primary-side switch of the power converter based on the transformed digital representation of the feedback signal.

Provided is an AD converter, including: an analog signal input circuit, configured to be input with an analog input signal, and output a first analog output signal based on the analog input signal and a second analog output signal based on the analog input signal at different timing; an integral circuit, configured to integrate the first analog output signal and the second analog output signal and output the first integral signal and the second integral signal; a predictive circuit, configured to predict an integral signal output after the output by the integral circuit based on the first integral signal and the second integral signal output by the integral circuit, and output a predictive integral signal; and a quantization circuit, configured to generate a digital signal with the predictive integral signal quantized.