Abstract: Devices and method for forming a switch including a heater layer including a first heater pad, a second heater pad, and a heater line connecting the first heater pad and the second heater pad, a phase change material (PCM) layer positioned in a same vertical plane as the heater line, and a floating spreader layer including a first portion positioned in the same vertical plane as the heater line and the PCM layer, in which the first portion has a first width that is less than or equal to a distance between proximate sidewalls of the first heater pad and the second heater pad.

Abstract: A semiconductor die package includes a high dielectric constant (high-k) dielectric layer over a device region of a first semiconductor die that is bonded with a second semiconductor die in a wafer on wafer (WoW) configuration. A through silicon via (TSV) structure may be formed through the device region. The high-k dielectric layer has an intrinsic negative charge polarity that provides a coupling voltage to modify the electric potential in the device region. In particular, the electron carriers in high-k dielectric layer attracts hole charge carriers in device region, which suppresses trap-assist tunnels that result from surface defects formed during etching of the recess for the TSV structure. Accordingly, the high-k dielectric layer described herein reduces the likelihood of (and/or the magnitude of) current leakage in semiconductor devices that are included in the device region of the first semiconductor die.

Abstract: A technique includes modifying a neural network model to sparsify the model. The model includes a plurality of kernel element weights, which are parameterized according to a plurality of dimensions. Modifying the model includes, in a given iteration of the plurality of iterations, training the model based on a structure regularization in which kernel element weights that share a dimension in common are removed as a group to create corresponding zero kernel elements in the model; and compressing the model to exclude zero kernel element weights from the model to prepare the model to be trained in another iteration.

Abstract: In one example, a device to process analog sensor data is described. For example, a device may include at least one analog sensor to generate a first set of analog voltage signals and a crossbar array including a plurality of memristors. In one example, the crossbar array is to receive an input vector of the first set of analog voltage signals, generate an output vector comprising a second set of analog voltage signals that is based upon a dot product of the input vector and a matrix comprising resistance values of the plurality of memristors, detect a pattern of the output vector, and activate a processor upon a detection of the pattern.

Abstract: Examples described herein relate to an analog front-end (AFE). The AFE includes a trans-impedance amplifier to receive an input current and generate a pair of the differential voltage signals based on the input current and a reference current. Further, the AFE includes a dynamic voltage slicer to receive the differential voltage signals at input terminals and supply digital voltages at output terminals. The dynamic voltage slicer includes a preamplifier to generate a pair of intermediate voltages based on the differential voltage signals sampled at a predetermined frequency. The dynamic voltage slicer also includes a voltage latch circuit coupled to the preamplifier, wherein the voltage latch circuit is to regenerate a pair of digital voltages based on the pair of the intermediate voltages. Moreover, the AFE includes a logic latch coupled to the dynamic voltage slicer to provide digital output states based on the pair of the digital voltages.

Abstract: Examples described herein relate to an analog front-end (AFE). The AFE includes a trans-impedance amplifier to receive an input current and generate a pair of the differential voltage signals based on the input current and a reference current. Further, the AFE includes a dynamic voltage slicer to receive the differential voltage signals at input terminals and supply digital voltages at output terminals. The dynamic voltage slicer includes a preamplifier to generate a pair of intermediate voltages based on the differential voltage signals sampled at a predetermined frequency. The dynamic voltage slicer also includes a voltage latch circuit coupled to the preamplifier, wherein the voltage latch circuit is to regenerate a pair of digital voltages based on the pair of the intermediate voltages. Moreover, the AFE includes a logic latch coupled to the dynamic voltage slicer to provide digital output states based on the pair of the digital voltages.

Abstract: A circuit includes a first power node configured to carry a first voltage having a first voltage level, a second power node configured to carry a second voltage having a second voltage level, an output node, and first and second cascode transistors coupled between the first power node and the output node and to each other at a node. A bias circuit uses the first and second cascode transistors to generate an output signal at the output node that transitions between the first voltage level and a third voltage level, and a delay circuit generates a transition in a first signal from one of the first or second voltage levels to the other of the first or second voltage levels, the transition having a time delay based on the output signal. A contending transistor couples the node to the second power node responsive to the first signal.

Abstract: Techniques and circuitry for wavelength monitor and control are disclosed herein. The disclosed wavelength monitor and control circuitry and techniques are designed to realize a multi-channel DWDM optical link by using a photonic receiver that dynamically adjusts resonant wavelengths of the microring drop filter (MDF), as needed. The wavelength monitor and control circuitry can monitor and control the resonant wavelengths of multiple MDFs for a DWDM silicon photonics receiver with minimum power and area overhead. In an embodiment, circuitry for an optical receiver comprises an MDF having resonant wavelength for multiple DWDM channels, and circuitry to control and monitor the resonant wavelength of the MDF in real-time and in manner that compensates for deviation between actual resonant wavelength of the MDF and the incident optical wavelength of the MDF.

Abstract: A circuit receives an input signal that switches between reference and first voltage levels, a power node carries a second voltage level, and a set of transistors is coupled between the power node and an output node. The second voltage level is a multiple of the first voltage level, and the multiple and a number of the transistors have a same value greater than two. A control signal circuit includes a level shifting circuit including a series of capacitive devices paired with latch circuits, a number of the pairs being one less than the value of the multiple, and, responsive to the input signal, outputs a control signal to a gate of a transistor of the first set of transistors closest to the power node, the control signal switching between the second voltage level and a third voltage level equal to the second voltage level minus the first voltage level.

Abstract: Examples described herein relate to reducing a magnitude of a supply voltage for a circuit element of an optical transmitter device. In some such examples, the circuit element is a driving element that is to receive a first electrical data signal and to provide a second electrical data signal to an optical element that is to provide an optical data signal. A testing element is to compare the optical data signal to the first electrical data signal to determine whether the optical transmitter device meets a performance threshold. When the device meets the performance threshold, a regulating element is to reduce a magnitude of the supply voltage of the driving element.

Abstract: A circuit includes a first power node configured to carry a first voltage having a first voltage level, a second power node configured to carry a second voltage having a second voltage level, an output node, and first and second cascode transistors coupled between the first power node and the output node and to each other at a node. A bias circuit uses the first and second cascode transistors to generate an output signal at the output node that transitions between the first voltage level and a third voltage level, and a delay circuit generates a transition in a first signal from one of the first or second voltage levels to the other of the first or second voltage levels, the transition having a time delay based on the output signal. A contending transistor couples the node to the second power node responsive to the first signal.

Abstract: A circuit receives an input signal that switches between reference and first voltage levels, a power node carries a second voltage level, and a set of transistors is coupled between the power node and an output node. The second voltage level is a multiple of the first voltage level, and the multiple and a number of the transistors have a same value greater than two. A control signal circuit includes a level shifting circuit including a series of capacitive devices paired with latch circuits, a number of the pairs being one less than the value of the multiple, and, responsive to the input signal, outputs a control signal to a gate of a transistor of the first set of transistors closest to the power node, the control signal switching between the second voltage level and a third voltage level equal to the second voltage level minus the first voltage level.

Abstract: A circuit includes a first power node configured to carry a first voltage having a first voltage level, an output node, a node coupled between the first power node and the output node, and a contending transistor coupled between the node and a second power node configured to carry a second voltage having a second voltage level. The circuit generates a signal at the output node that ranges between the first voltage level and a third voltage level, the contending transistor couples the node with the second power node responsive to the signal, a difference between the first voltage level and the second voltage level has a first magnitude, a difference between the first voltage level and the third voltage level has a second magnitude, and the second magnitude is a multiple of the first magnitude having a value greater than one.

Abstract: A level-shifting circuit includes an input device configured to receive an input signal capable of switching between a reference voltage level and a first voltage level, and a set of capacitive devices paired in series with latch circuits. A first capacitive device of the set is coupled with an output of the input device, and each capacitive device and latch circuit pair is configured to upshift a corresponding received signal by an amount equal to a difference between the first voltage level and the reference voltage level.

Abstract: Examples described herein relate to reducing a magnitude of a supply voltage for a circuit element of an optical transmitter device. In some such examples, the circuit element is a driving element that is to receive a first electrical data signal and to provide a second electrical data signal to an optical element that is to provide an optical data signal. A testing element is to compare the optical data signal to the first electrical data signal to determine whether the optical transmitter device meets a performance threshold. When the device meets the performance threshold, a regulating element is to reduce a magnitude of the supply voltage of the driving element.

Abstract: An example system includes an optical modulator and a multiplexing controller. The modulator includes a data bus for receiving at least one data signal, a plurality of multiplexers and a plurality of modulating segments. Each multiplexer is coupled to the data bus to receive at least one data signal and to output a multiplexed signal. Each modulating segment may receive the multiplexed signal from one of the plurality of multiplexers and modulate the multiplexed signal using an optical input. The multiplexing controller may be in communication with the plurality of multiplexers and may configure each of the plurality of multiplexers in accordance with a selected modulation type.

Abstract: A current value of a first pixel and/or a current value of a second pixel of a display are adjusted until a value of a current difference is within a predetermined range. The current value of the first pixel corresponds to a brightness level of the first pixel. The current value of the second pixel corresponds to a brightness level of the second pixel. Adjusting the current value of the first pixel involves adjusting a threshold voltage value of a transistor of the first pixel. Adjusting the current value of the second pixel involves adjusting a threshold voltage value of a transistor of the second pixel.

Abstract: In example implementations, a method executed by a processor is provided. The method receives a simulated photonic data input based on a theoretical photonic design that meets a target specification. A complementary metal-oxide semiconductor (CMOS) circuit design is designed based on the simulated photonic data input using a pre-layout simulation. An experimental photonic data input based on a fabricated photonics device that meets the target specification is received. The CMOS circuit is designed based on the experimental photonic data input using a post-layout simulation. A physical circuit CMOS circuit design and a layout that includes detailed physical dimensions associated with the physical CMOS circuit design that is based on the pre-layout and the post-layout are transmitted to a CMOS foundry.

Abstract: A technique includes modifying a neural network model to sparsify the model. The model includes a plurality of kernel element weights, which are parameterized according to a plurality of dimensions. Modifying the model includes, in a given iteration of the plurality of iterations, training the model based on a structure regularization in which kernel element weights that share a dimension in common are removed as a group to create corresponding zero kernel elements in the model; and compressing the model to exclude zero kernel element weights from the model to prepare the model to be trained in another iteration.

Abstract: A circuit includes a startup circuit to provide a charging signal to initiate startup of a reference circuit. The startup circuit includes a detector circuit having a detector current path control, a level shifter having a level shifter current path control, and a charger circuit having a charger current path control. Each of the detector current path control, the level shifter current path control, and the charger circuit current path control enable current flow in the startup circuit when the charger turn-on signal is in the on-state and disable the current flow in the startup circuit when the charger turn-on signal is in the off state.