Abstract: Prior knowledge of access pattern is leveraged to improve energy dissipation for general matrix operations. This improves memory access energy for a multitude of applications such as image processing, deep neural networks, and scientific computing workloads, for example. In some embodiments, prior knowledge of access pattern allows for burst read and/or write operations. As such, burst mode solution can provide energy savings in both READ (RD) and WRITE (WR) operations. For machine learning or inference, the weight values are known ahead in time (e.g., inference operation), and so the unused bytes in the cache line are exploited to store a sparsity map that is used for disabling read from either upper or lower half of the cache line, thus saving dynamic capacitance.

Abstract: An apparatus, system, and method for register file circuits are provided. A register file circuit can include a first write bit line (WBL), a first P-channel metal oxide semiconductor (PMOS) transistor including a source coupled to the WBL, a first inverter including an input coupled to a drain of the first PMOS transistor, a second PMOS transistor including a source coupled to an output of the first inverter, and a second WBL (WBLB) coupled to a drain of the second PMOS transistor. 1R1W register file and 2R1W register file designs are provided.

Abstract: A distributed and scalable all-digital LDO (D-DLDO) voltage regulator allowing rapid scaling across technology nodes. The distributed DLDO includes many tillable DLDO units regulating a single supply voltage with a shared power distribution network (PDN). The D-DLDO includes an all-digital proportional-integral-derivative (PID) controller that receives a first code indicative of a voltage behavior on a power supply rail. A droop detector is provided to compare the first code with a threshold to determine a droop event, wherein information about the droop event is provided to the PID controller, wherein the PID controller generates a second code according to the first code and the information about the droop event. The DLDO includes a plurality of power gates that receive the second code.

Abstract: Embodiments herein relate to scaling of Static Random Access Memory (SRAM) cells. An SRAM cell include nMOS transistors on one level above pMOS transistors on a lower level. Transistors on the two levels can have overlapping footprints to save space. Additionally, the SRAM cell can use pMOS access transistors in place of nMOS access transistors to allow reuse of areas of the cell which would otherwise be used by the nMOS access transistors. In one approach, gate interconnects are provided in these areas, which have an overlapping footprint with underlying pMOS access transistors to save space. The SRAM cells can be connected to bit lines and word lines in overhead and/or bottom metal layers. In another aspect, SRAM cells of a column are connected to bit lines in an overlying M0 metal layer and an underlying BM0 metal layers to reduce capacitance.

Abstract: A system includes a compute circuit that preemptively performs a computation on a data word before receiving an indication of data errors from an error checking and correction (ECC) circuit. The ECC circuit reads the data word from a memory array and performs error detection and error correction on the data word. The compute circuit reads the data word and performs the computation on the data word to generate an output value, without waiting for the ECC circuit to check and correct the data word. In response to error detection in the data word by the ECC circuit, the compute circuit delays outputting the output value until correction of the output value in accordance with the error detection by the ECC circuit.

Abstract: An ultra-deep compute Static Random Access Memory (SRAM) with high compute throughput and multi-directional data transfer capability is provided. Compute units are placed in both horizontal and vertical directions to achieve a symmetric layout while enabling communication between the compute units. An SRAM array supports simultaneous read and write to the left and right section of the same SRAM subarray by duplicating pre-decoding logic inside the SRAM array. This allows applications with non-overlapping read and write address spaces to have twice the bandwidth as compared to a baseline SRAM array.

Abstract: An apparatus is provided which comprises: a plurality of devices coupled to an input power supply rail and an output power supply rail; a first circuitry coupled to the plurality of devices, wherein the first circuitry is to turn on or off one or more devices of the plurality according to a control; and a second circuitry coupled to the first circuitry, wherein the second circuitry comprises an all-digital proportional-derivative mechanism to generate the control according to a digital representation of voltage on the output power supply rail.

Abstract: Described is a controller that provides in-situ state retention using a closed loop global retention clamp. The controller addresses di/dt and reliability constraints using an adaptive scheme where steps with smaller current are quickly changed whereas steps with larger current are changed slowly. The loop controller of a voltage regulator is modified for controlling not only retention Vmin during a low power state (e.g., C1LP), but also to control fast wake up the low power state (e.g., from C1LP and from C6).

Abstract: An all-digital voltage monitor (ADVM) generates a multi-bit output code that changes in proportion to a voltage being monitored, by leveraging the voltage impact on a gate delay. ADVM utilizes a simple delay chain, which receives a clock-cycle-long pulse every clock cycle, such that the monitored supply voltage is sampled for one full cycle every cycle. The outputs of all delay cells of the delay chain collectively represents a current voltage state as a digital thermometer code. In AVDM, a voltage droop event thus results in a decrease in the output code from a nominal value, while an overshoot results in an increase in the output code.

Abstract: Embodiments include apparatuses, methods, and systems to implement a multi-read and/or multi-write process with a set of memory cells. The set of memory cells may be multiplexed with a same sense amplifier. As part of a multi-read process, a memory controller coupled to a memory circuit may precharge the bit lines associated with the set of memory cells, provide a single assertion of a word line signal on the word line, and then sequentially read data from the set of memory cells (using the sense amplifier) based on the precharge and the single assertion of the word line signal. Additionally, or alternatively, a multi-write process may be performed to sequentially write data to the set of memory cells based on one precharge of the associated bit lines. Other embodiments may be described and claimed.

Abstract: An all-digital voltage monitor (ADVM) generates a multi-bit output code that changes in proportion to a voltage being monitored, by leveraging the voltage impact on a gate delay. ADVM utilizes a simple delay chain, which receives a clock-cycle-long pulse every clock cycle, such that the monitored supply voltage is sampled for one full cycle every cycle. The outputs of all delay cells of the delay chain collectively represents a current voltage state as a digital thermometer code. In AVDM, a voltage droop event thus results in a decrease in the output code from a nominal value, while an overshoot results in an increase in the output code.

Abstract: A distributed and scalable all-digital LDO (D-DLDO) voltage regulator allowing rapid scaling across technology nodes. The distributed DLDO includes many tillable DLDO units regulating a single supply voltage with a shared power distribution network (PDN). The D-DLDO includes an all-digital proportional-integral-derivative (PID) controller that receives a first code indicative of a voltage behavior on a power supply rail. A droop detector is provided to compare the first code with a threshold to determine a droop event, wherein information about the droop event is provided to the PID controller, wherein the PID controller generates a second code according to the first code and the information about the droop event. The DLDO includes a plurality of power gates that receive the second code.

Abstract: Prior knowledge of access pattern is leveraged to improve energy dissipation for general matrix operations. This improves memory access energy for a multitude of applications such as image processing, deep neural networks, and scientific computing workloads, for example. In some embodiments, prior knowledge of access pattern allows for burst read and/or write operations. As such, burst mode solution can provide energy savings in both READ (RD) and WRITE (WR) operations. For machine learning or inference, the weight values are known ahead in time (e.g., inference operation), and so the unused bytes in the cache line are exploited to store a sparsity map that is used for disabling read from either upper or lower half of the cache line, thus saving dynamic capacitance.

Abstract: A system includes a compute circuit that preemptively performs a computation on a data word before receiving an indication of data errors from an error checking and correction (ECC) circuit. The ECC circuit reads the data word from a memory array and performs error detection and error correction on the data word. The compute circuit reads the data word and performs the computation on the data word to generate an output value, without waiting for the ECC circuit to check and correct the data word. In response to error detection in the data word by the ECC circuit, the compute circuit delays outputting the output value until correction of the output value in accordance with the error detection by the ECC circuit.

Abstract: Embodiments include apparatuses, methods, and systems to implement a multi-read and/or multi-write process with a set of memory cells. The set of memory cells may be multiplexed with a same sense amplifier. As part of a multi-read process, a memory controller coupled to a memory circuit may precharge the bit lines associated with the set of memory cells, provide a single assertion of a word line signal on the word line, and then sequentially read data from the set of memory cells (using the sense amplifier) based on the precharge and the single assertion of the word line signal. Additionally, or alternatively, a multi-write process may be performed to sequentially write data to the set of memory cells based on one precharge of the associated bit lines. Other embodiments may be described and claimed.

Abstract: An apparatus, vision processing unit, and method are provided for clustering motion events in a content addressable memory. A motion event is received including coordinates in an image frame that have experienced a change and a timestamp of the change. A determination is made as to whether determine whether there is a valid entry in the memory having coordinates within a predefined range of coordinates included in the motion event. In response to a determination that there is the valid entry having the coordinates within the predefined range of coordinates included in the motion event, write to the valid entry the coordinates and the timestamp in the motion event.

Abstract: A spike sent from a first artificial neuron in a spiking neural network (SNN) to a second artificial neuron in the SNN is identified, with the spike sent over a particular artificial synapse in the SNN. The membrane potential of the second artificial neuron at a particular time step, corresponding to sending of the spike, is compared to a threshold potential, where the threshold potential is set lower than a firing potential of the second artificial neuron. A change to the synaptic weight of the particular artificial synapse is determined based on the spike, where the synaptic weight is to be decreased if the membrane potential of the second artificial neuron is lower than the threshold potential at the particular time step and the synaptic weight is to be increased if the membrane potential of the second artificial neuron is higher than the threshold potential at the particular time step.

Abstract: An all-digital voltage monitor (ADVM) generates a multi-bit output code that changes in proportion to a voltage being monitored, by leveraging the voltage impact on a gate delay. ADVM utilizes a simple delay chain, which receives a clock-cycle-long pulse every clock cycle, such that the monitored supply voltage is sampled for one full cycle every cycle. The outputs of all delay cells of the delay chain collectively represents a current voltage state as a digital thermometer code. In AVDM, a voltage droop event thus results in a decrease in the output code from a nominal value, while an overshoot results in an increase in the output code.

Abstract: Embodiments include apparatuses, methods, and systems to implement a multi-read and/or multi-write process with a set of memory cells. The set of memory cells may be multiplexed with a same sense amplifier. As part of a multi-read process, a memory controller coupled to a memory circuit may precharge the bit lines associated with the set of memory cells, provide a single assertion of a word line signal on the word line, and then sequentially read data from the set of memory cells (using the sense amplifier) based on the precharge and the single assertion of the word line signal. Additionally, or alternatively, a multi-write process may be performed to sequentially write data to the set of memory cells based on one precharge of the associated bit lines. Other embodiments may be described and claimed.

Abstract: A processor or integrated circuit includes a memory to store weight values for a plurality neuromorphic states and a circuitry coupled to the memory. The circuitry is to detect an incoming data signal for a pre-synaptic neuromorphic state and initiate a time window for the pre-synaptic neuromorphic state in response to detecting the incoming data signal. The circuitry is further to, responsive to detecting an end of the time window: retrieve, from the memory, a weight value for a post-synaptic neuromorphic state for which an outgoing data signal is generated during the time window, the post-synaptic neuromorphic state being a fan-out connection of the pre-synaptic neuromorphic state; perform a causal update to the weight value, according to a learning function, to generate an updated weight value; and store the updated weight value back to the memory.