Abstract: The present invention relates to a system comprising a pair of System-on-Chip dies that includes through-silicon vias (TSVs) coupling the front side and the back side. A first fanout wafer with first fanout layers is coupled to the front side of the System-on-Chip dies and second fanout wafer with second fanout layers is coupled to the TSVs on the back side of the System-on-Chip dies. A carrier wafer is bonded to the second fanout wafer, collectively forming a 3D double-sided fanout structure that approximately doubles the available die-to-die routing bandwidth.

Abstract: An apparatus for receiving data, wherein the apparatus is configured to: receive a multiplexed data stream over an optical link, the multiplexed data stream comprising: a first data stream sent from a first transmitter, the first data stream having a minimum time interval between condition changes; a second data stream sent from a second transmitter, the second data stream having the same minimum time interval between condition changes, wherein the second data stream is delayed by a fraction of the minimum time interval relative to the first data stream.

Abstract: A method of transmission, the method comprising: sending, in two or more groups of channels, data and at least one clock signal over an optical link to a receiver, wherein each group in the two or more groups of channels comprises: at least one channel for sending at least one data packet of the data; and a channel for sending a clock signal of the at least one clock signal, wherein the clock signal is synchronized with the at least one data packet in the group of channels.

Abstract: A transmission method comprising: encoding a message into error correction coding blocks, each comprising a different portion of the payload bits of the message and a respective plurality of error correction bits; and applying an interleaving map to error correction coding blocks, such that for each error correction coding block, different bits of the error correction coding block are mapped to different channels, each channel thus being used to transmit bits from a respective selection of multiple of the error correction coding blocks. Thus, for at least one channel failure scenario in which at least one of the plurality of channel fails leaving a plurality of remaining channels, enough bits will remain on the remaining channels to enable correction of the message based on the error correction coding.

Abstract: A transmission method comprising: encoding a message into error correction coding blocks, each comprising a different portion of the payload bits of the message and a respective plurality of error correction bits; and applying an interleaving map to error correction coding blocks, such that for each error correction coding block, different bits of the error correction coding block are mapped to different channels, each channel thus being used to transmit bits from a respective selection of multiple of the error correction coding blocks. Thus, for at least one channel failure scenario in which at least one of the plurality of channel fails leaving a plurality of remaining channels, enough bits will remain on the remaining channels to enable correction of the message based on the error correction coding.

Abstract: A domain control circuit includes a power regulator to supply power for a first domain on a power rail and a sequencing circuit to control the power regulator, and a clock gate signal to activate the domain. The sequencing circuit receives a domain control signal to control activation and deactivation of the domain. The domain control circuit deactivates the clock gate signal to the domain after controlling the power regulator to supply power for the domain on a power rail. In this manner, a voltage droop in a supply voltage on a power rail is reduced. In some examples, the clock gate signal to the domain is deactivated after a voltage increase on the power rail. In some examples, the power regulator includes a plurality of parallel regulator circuits and a regulator control circuit to determine a number of the parallel regulator circuits to be activated to power the domain.

Abstract: A domain control circuit includes a power regulator to supply power for a first domain on a power rail and a sequencing circuit to control the power regulator, and a clock gate signal to activate the domain. The sequencing circuit receives a domain control signal to control activation and deactivation of the domain. The domain control circuit deactivates the clock gate signal to the domain after controlling the power regulator to supply power for the domain on a power rail. In this manner, a voltage droop in a supply voltage on a power rail is reduced. In some examples, the clock gate signal to the domain is deactivated after a voltage increase on the power rail. In some examples, the power regulator includes a plurality of parallel regulator circuits and a regulator control circuit to determine a number of the parallel regulator circuits to be activated to power the domain.

Abstract: A digital signal processor includes analog to digital converters to convert an analog voltage to digital voltage in unit intervals of an analog signal. A decision feedback equalizer (DFE) determines a first level of a digital sum of a digital voltage in a first UI and digital voltages of adjacent UIs (taps). The DFE identifies predetermined sequences of levels of consecutive UIs that include the first level and selects one of the predetermined sequences to decode digital data encoded in the analog signal in the UI. The DSP may be programmable to include taps from UIs that may affect the first UI. The predetermined sequences may include levels of the digital sums of consecutive UIs of the analog signal. The predetermined sequences may be identified in a look-up table based on the first level.

Abstract: A system for interconnecting a plurality of computing nodes includes a plurality of optical circuit switches and a plurality of electrical circuit switches. A first network stage comprises a first plurality of circuit switches selected from among the plurality of optical circuit switches and the plurality of electrical circuit switches. Each computing node among the plurality of computing nodes is optically coupled to at least one of the first plurality of circuit switches. A second network stage comprises a second plurality of circuit switches selected from among the plurality of optical circuit switches and the plurality of electrical circuit switches. Each circuit switch among the first plurality of circuit switches is optically coupled to each circuit switch among the second plurality of optical circuit switches.

Abstract: A phase-locked loop (PLL) circuit generates an analog signal in phase-lock with a reference signal at a reference frequency. The PLL circuit includes a charge pump circuit, a loop filter circuit, a feedback divider, and a voltage controlled oscillator (VCO). The charge pump circuit charges a sample capacitor of the loop filter circuit to a sample voltage based on a phase difference between the generated analog signal and the reference signal. The loop filter circuit stores the sample voltage as a proportional control voltage in a hold capacitor to reduce or avoid ripple in the control voltage that causes jitter in the analog signal. The loop filter circuit also provides the sample voltage to an integral component circuit comprising a comparator and digital accumulator producing an integral control. The VCO generates the analog signal at a frequency based on the proportional control voltage and the integral control voltage.

Abstract: Analog calibration (ACAL) circuits supporting iterative measurement of an input signal from a measured circuit, and related methods are disclosed. The ACAL circuit includes a voltage reference generation circuit and a comparator circuit. The voltage reference generation circuit is configured to provide an input reference voltage. The comparator circuit is configured to compare the input reference voltage to an input circuit voltage of a measured circuit and generate a digital measurement signal based on the comparison. To provide for the ACAL circuit to more precisely measure the input circuit voltage, the voltage reference generation circuit is programmable and is configured to a generate the input reference voltage based on a programmed reference voltage selection. In this manner, the ACAL circuit can be used to measure the input circuit voltage in an iterative manner based on different programmed input reference voltages for a more precise measurement of the input circuit voltage.

Abstract: A phase-locked loop (PLL) circuit generates an analog signal in phase-lock with a reference signal at a reference frequency. The PLL circuit includes a charge pump circuit, a loop filter circuit, a feedback divider, and a voltage controlled oscillator (VCO). The charge pump circuit charges a sample capacitor of the loop filter circuit to a sample voltage based on a phase difference between the generated analog signal and the reference signal. The loop filter circuit stores the sample voltage as a proportional control voltage in a hold capacitor to reduce or avoid ripple in the control voltage that causes jitter in the analog signal. The loop filter circuit also provides the sample voltage to an integral component circuit comprising a comparator and digital accumulator producing an integral control. The VCO generates the analog signal at a frequency based on the proportional control voltage and the integral control voltage.

Abstract: A system for interconnecting a plurality of computing nodes includes a plurality of optical circuit switches and a plurality of electrical circuit switches. A first network stage comprises a first plurality of circuit switches selected from among the plurality of optical circuit switches and the plurality of electrical circuit switches. Each computing node among the plurality of computing nodes is optically coupled to at least one of the first plurality of circuit switches. A second network stage comprises a second plurality of circuit switches selected from among the plurality of optical circuit switches and the plurality of electrical circuit switches. Each circuit switch among the first plurality of circuit switches is optically coupled to each circuit switch among the second plurality of optical circuit switches.

Abstract: Analog calibration (ACAL) circuits supporting iterative measurement of an input signal from a measured circuit, and related methods are disclosed. The ACAL circuit includes a voltage reference generation circuit and a comparator circuit. The voltage reference generation circuit is configured to provide an input reference voltage. The comparator circuit is configured to compare the input reference voltage to an input circuit voltage of a measured circuit and generate a digital measurement signal based on the comparison. To provide for the ACAL circuit to more precisely measure the input circuit voltage, the voltage reference generation circuit is programmable and is configured to a generate the input reference voltage based on a programmed reference voltage selection. In this manner, the ACAL circuit can be used to measure the input circuit voltage in an iterative manner based on different programmed input reference voltages for a more precise measurement of the input circuit voltage.

Abstract: A system for interconnecting a plurality of computing nodes includes a plurality of optical circuit switches and a plurality of electrical circuit switches. A first network stage comprises a first plurality of circuit switches selected from among the plurality of optical circuit switches and the plurality of electrical circuit switches. Each computing node among the plurality of computing nodes is optically coupled to at least one of the first plurality of circuit switches. A second network stage comprises a second plurality of circuit switches selected from among the plurality of optical circuit switches and the plurality of electrical circuit switches. Each circuit switch among the first plurality of circuit switches is optically coupled to each circuit switch among the second plurality of optical circuit switches.

Abstract: A system for interconnecting a plurality of computing nodes includes a plurality of optical circuit switches and a plurality of electrical circuit switches. A first network stage comprises a first plurality of circuit switches selected from among the plurality of optical circuit switches and the plurality of electrical circuit switches. Each computing node among the plurality of computing nodes is optically coupled to at least one of the first plurality of circuit switches. A second network stage comprises a second plurality of circuit switches selected from among the plurality of optical circuit switches and the plurality of electrical circuit switches. Each circuit switch among the first plurality of circuit switches is optically coupled to each circuit switch among the second plurality of optical circuit switches.

Abstract: The present disclosure is generally directed to a method and apparatus to communicate data between two or more semiconductor devices. In an embodiment, a method includes synchronizing a master device with a slave device, where the master device includes a semiconductor device. Synchronizing includes transmitting a first synchronization marker data pattern via a first serial interface from the master device at a first time, and receiving a second synchronization marker data pattern via a second serial interface at the master device at a second time in response to transmitting the first synchronization marker data pattern. Synchronizing also includes determining, based at least in part on the first time and the second time, a third time when a reply is to be received by the master device in response to a request transmitted from the master device to the slave device.

Abstract: The present disclosure is generally directed to a method and apparatus to communicate data between two or more semiconductor devices. In an embodiment, a method includes synchronizing a master device with a slave device, where the master device includes a semiconductor device. Synchronizing includes transmitting a first synchronization marker data pattern via a first serial interface from the master device at a first time, and receiving a second synchronization marker data pattern via a second serial interface at the master device at a second time in response to transmitting the first synchronization marker data pattern. Synchronizing also includes determining, based at least in part on the first time and the second time, a third time when a reply is to be received by the master device in response to a request transmitted from the master device to the slave device.

Abstract: The present disclosure is generally directed to a method of communicating data between two or more semiconductor devices. Serial interfaces using the method have a reduction in latency compared to conventional serial interfaces. The method enables features needed for a serial interface, such as limited run lengths and recognizable data boundaries to establish alignment. In addition, a method for synchronizing two or more semiconductor devices through serial interfaces has been presented. This is done by passing a marker data pattern through the system.

Abstract: A memory device and system are disclosed that may include a serial data interface, a serial address interface, and a reference clock interface. The reference clock interface is configured to receive a signal from a reference clock source that provides a reference clock signal to a memory control device. The serial interfaces are coupled to other memory devices or memory control devices. A method of transferring data within a memory system using serial interfaces is also disclosed. The method includes performing clock multiplication on a reference clock to provide a multiplied clock, using the multiplied clock to serialize and transmit data onto a serial interface, recovering data from the seal interface, using the reference clock to determine an initial frequency for use by clock and data recovery module, using the data recovered from the serial interface to determine a phase and final frequency of a recovered clock, and using the recovered clock to de-serialize received serial data into parallel data words.