Abstract: Embodiments are directed towards a configurable accelerator framework device that includes a stream switch and a plurality of convolution accelerators. The stream switch has a plurality of input ports and a plurality of output ports. Each of the input ports is configurable at run time to unidirectionally pass data to any one or more of the output ports via a stream link. Each one of the plurality of convolution accelerators is configurable at run time to unidirectionally receive input data via at least two of the plurality of stream switch output ports, and each one of the plurality of convolution accelerators is further configurable at run time to unidirectionally communicate output data via an input port of the stream switch.

Abstract: A charge pump circuit includes a first charge pump stage circuit coupled in series with a second charge pump stage circuit. A discharge circuit operates to discharge the charge pump circuit. The discharge circuit includes: a first switched circuit coupled to a first output of the first charge pump stage circuit and configured, when actuated, to discharge the first output; and a second switched circuit coupled to a second output of the second charge pump stage circuit and configured, when actuated, to discharge the second output. A discharge control circuit actuates the first switched discharge circuit to discharge the first output and then, after the first output is fully discharged, actuates the second switched discharge circuit to discharge the second output.

Abstract: A phase lock loop (PLL) includes an input comparison circuit configured to compare a reference signal to a divided feedback signal and generate at least one charge pump control signal based thereupon. A charge pump generates a charge pump output signal in response to the at least one charge pump control signal. A loop filter is coupled to receive and filter the charge pump output signal to produce an oscillator control signal. An oscillator generates an output signal in response to the oscillator control signal, with the output signal divided by a divisor using divider circuitry to produce the divided feedback signal. Divisor generation circuitry is configured to change the divisor over time so that a frequency of the divided feedback signal changes from a first frequency to a second frequency over time.

Abstract: A method of corrupting contents of a memory array includes asserting a signal at a reset node to thereby cause starving of current supply to the memory array, and selecting bit lines and complementary bit lines associated with desired columns of the memory array that contain memory cells to have their contents corrupted. For each desired column, a logic state of its bit line and complementary bit line are forced to a same logic state. Each word line associated with desired rows of the memory array that contains memory cells to have their contents corrupted is simultaneously asserted, and then simultaneously deasserted to thereby place each memory cell to have its contents corrupted into a metastable state during a single clock cycle.

Abstract: An in-memory computation circuit includes a memory array with SRAM cells connected in rows by word lines and in columns by bit lines. Body bias nodes of the transistors in each SRAM cell are biased by a modulated body bias voltage. A row controller circuit simultaneously actuates word lines in parallel for an in-memory compute operation. A column processing circuit processes analog voltages developed on the bit lines in response to the simultaneous actuation to generate a decision output for the in-memory compute operation. A voltage generator circuit switches the modulated body bias voltage from a non-negative voltage level to a negative voltage level during the simultaneous actuation. The negative voltage level is adjusted dependent on integrated circuit process and/or temperature conditions in order to optimize protection against unwanted memory cell data flip.

Abstract: A method of memory reset includes precharging bit lines of a memory array, asserting a signal at a reset node to remove the precharge voltage, and selecting write drivers associated with the bit lines associated with columns of the memory array that contain memory cells to be reset, with the assertion of the signal at the reset node also resulting in application of desired logic states to inputs of the selected write drivers to cause those selected write drivers to change a logic state of the bit lines associated with those write drivers. The method continues with asserting each word line associated with a row of the memory that contains memory cells to be reset to write desired logic states to all of the memory cells of the columns and rows of the memory to be reset during a single clock cycle, and then deasserting those word lines.

Abstract: Disclosed herein is logic circuitry and techniques for operation that hardware to enable the construction of first-in-first-out (FIFO) buffers from latches while permitting stuck-at-1 fault testing for the enable pin of those latches, as well as testing the data path at individual points through the FIFO buffer.

Abstract: An integrated circuit die includes memory sectors, each memory sector including a memory array. The die includes a voltage regulator with a first transistor driven by an output voltage to thereby generate a gate voltage, the output voltage being generated based upon a difference between a constant current and a leakage current. A selection circuit selectively couples the gate voltage to a selected one of the plurality of memory sectors. A leakage detector circuit drives a second transistor with the output voltage to thereby generate a copy voltage based upon a difference between a variable current and a replica of the constant current, increases the variable current in response to the copy voltage being greater than the gate voltage, and asserts a leakage detection signal in response to the copy voltage being less than the gate voltage, the leakage detection signal indicating excess leakage within the memory array.

Abstract: A circuit includes a memory array with SRAM cells connected in rows by word lines and in columns by bit lines. A row controller circuit simultaneously actuates, through a word line driver circuit for each row, word lines in parallel for an in-memory compute operation. A column processing circuit processes analog voltages developed on the bit lines in response to the simultaneous actuation to generate a decision output for the in-memory compute operation. A bit line clamping circuit includes a sensing circuit that compares the analog voltages on a given pair of bit lines to a threshold voltage. A voltage clamp circuit is actuated in response to the comparison to preclude the analog voltages on the given pair of bit lines from decreasing below a clamping voltage level.

Abstract: An in-memory computation circuit includes a memory array with SRAM cells connected in rows by word lines and in columns by bit lines. Each row includes a word line drive circuit powered by an adaptive supply voltage. A row controller circuit simultaneously actuates word lines in parallel for an in-memory compute operation. A column processing circuit processes analog voltages developed on the bit lines in response to the simultaneous actuation to generate a decision output for the in-memory compute operation. A voltage generator circuit generates the adaptive supply voltage for powering the word line drive circuits during the simultaneous actuation. A level of the adaptive supply voltage is modulated dependent on integrated circuit process and/or temperature conditions in order to optimize word line underdrive performance and inhibit unwanted memory cell data flip.

Abstract: A circuit includes a memory array with SRAM cells connected in rows by word lines and in columns by bit lines. A row controller circuit simultaneously actuates, through a word line driver circuit for each row, word lines in parallel for an in-memory compute operation. A column processing circuit processes analog voltages developed on the bit lines in response to the simultaneous actuation to generate a decision output for the in-memory compute operation. A bit line precharge circuit generates a precharge voltage for application to each pair of bit lines. The precharge voltage has a first voltage level (not greater than a positive supply voltage for the SRAM cells) when the memory array is operating in a data read/write mode. The precharge voltage has a second voltage level (greater than the first voltage level) in advance of the simultaneous actuation of the word lines for the in-memory compute operation.

Abstract: SRAM cells are connected in columns by bit lines and connected in rows by first and second word lines coupled to first and second data storage sides of the SRAM cells. First the first word lines are actuated in parallel and then next the second word lines are actuated in parallel in first and second phases, respectively, of an in-memory compute operation. Bit line voltages in the first and second phases are processed to generate an in-memory compute operation decision. A low supply node reference voltage for the SRAM cells is selectively modulated between a ground voltage and a negative voltage. The first data storage side receives the negative voltage and the second data storage side receives the ground voltage during the second phase. Conversely, the second data storage side receives the negative voltage and the first data storage side receives the ground voltage during the first phase.

Abstract: A method to bypass a voltage regulator of a system on a chip (SOC) comprising powering a first power domain using a voltage regulator; powering a second power domain using the voltage regulator; coupling a third power domain with an external voltage source; raising an external voltage supply from the external voltage source above a threshold level of the voltage regulator; coupling the first second power domains to the external voltage source; turning OFF the voltage regulator of the SOC after coupling the first power domain of the SOC and the second power domain of the SOC to the external voltage source; and powering the first power domain of the SOC, the second power domain of the SOC, and the third power domain of the SOC with the external voltage source, the external voltage source bypassing the voltage regulator.

Abstract: The present disclosure is directed to arranging user data memory cells and test memory cells in a configurable memory array that can perform both differential and single ended read operations during memory start-up and normal memory use, respectively. Different arrangements of the user data memory cells and the test memory cells in the memory array result in increased effectiveness of memory array, in terms of area optimization, memory read accuracy and encryption for data security.

Abstract: A method includes receiving, at a master agent, announcements from candidate consumer agents indicating the presence of the candidate consumer agents. Each announcement includes display parameters for a display of the corresponding candidate consumer agent. The method further includes receiving at the master agent content parameters from a producer agent, the content parameters defining characteristics of content to be provided by the consumer agent. A mosaic screen is configured based on the received announcements and the content parameters. This configuring of the mosaic screen includes selecting ones of the consumer agents for which an announcement was received and generating content distribution parameters based on the content parameters and the display parameters of the selected ones of the consumer agents. The generated content distribution parameters are provided to the consumer agent.

Abstract: A time-interleaved analog to digital converter (TI-ADC) includes a first sub-ADC configured to sample and convert an input analog signal to generate a first digital signal and a second sub-ADC configured to sample and convert said input analog signal to generate a second digital signal. Sampling by the second sub-ADC occurs with a time skew mismatch. A multiplexor interleaves the first and second digital signals to generate a third digital signal. A time skew mismatch error determination circuit processes the first and second digital signals to generate a time error corresponding to the time skew mismatch. A slope value of said third digital signal is determined and multiplied by the time error to generate a signal error. The signal error is summed with the third digital signal to generate a digital output signal which eliminates the error due to the time skew mismatch. This correction is performed in real time.

Abstract: A serial peripheral interface (SPI) device includes a serial clock (SCK) pad receiving a serial clock, first and second Schmitt triggers directly electrically connected to the SCK pad to selectively respectively generate first and second clocks in response to rising and falling edges of the serial clock, first and second flip flops clocked by the first and second clocks to output bits of data to a data node, a multiplexer having an input coupled to the data node and an output coupled to driving circuitry, and driving circuitry transmitting data via a master-in-slave-out (MISO) pad.

Abstract: A memory array arranged in multiple columns and rows. Computation circuits that each calculate a computation value from cell values in a corresponding column. A column multiplexer cycles through multiple data lines that each corresponds to a computation circuit. Cluster cycle management circuitry determines a number of multiplexer cycles based on a number of columns storing data of a compute cluster. A sensing circuit obtains the computation values from the computation circuits via the column multiplexer as the column multiplexer cycles through the data lines. The sensing circuit combines the obtained computation values over the determined number of multiplexer cycles. A first clock may initiate the multiplexer to cycle through its data lines for the determined number of multiplexer cycles, and a second clock may initiate each individual cycle. The multiplexer or additional circuitry may be utilized to modify the order in which data is written to the columns.

Abstract: An integrated circuit includes a pulse width modulator. The pulse width modulator includes a multiplexer that receives a plurality of data delay signals. Each of the data delay signals is based on a data signal and a respective clock phase signal. The multiplexer includes a first multiplexer stage and a second multiplexer stage. The first multiplexer stage receives all of the data delay signals and has a relatively large delay. The second multiplexer stage receives to output signals from the first multiplexer stage and has a relatively small delay. The second multiplexer stage outputs a pulse width modulation signal that can have a pulse width corresponding to the offset between two adjacent clock phase signals.

Abstract: Data structure and microcontroller architecture performing binary multiply-accumulate operations using multiple partial copies of weights. Destination-register location, source-register location, and weight-register location are received. Using the weight-register location, a sub-set of the weight bits is copied a select number of times based on a filter index value that is received. Each copy of the sub-set of weights is executed in parallel. Using the source-register location, a sub-set of the input bits is selected based on the size of the sub-set of weights, wherein the sub-set of input bits is shifted one bit from a previous sub-set of input bits. XOR operation is performed on each corresponding bit in the copy of the sub-set of weights with each corresponding bit in the selected sub-set of input bits. In a corresponding destination sub-location, output of each XOR operation is aggregated with each other and with current value of the corresponding destination sub-location.