Abstract: A memory circuit has compute-in-memory circuitry that enables a multiply-accumulate (MAC) operation based on shared charge. Row access circuitry drives multiple rows of a memory array to multiply a first data word with a second data word stored in the memory array. The row access circuitry drives the multiple rows based on the bit pattern of the first data word. Column access circuitry drives a column of the memory array when the rows are driven. Accessed rows discharge the column line in an accumulative fashion. Sensing circuitry can sense voltage on the column line. A processor in the memory circuit computes a MAC value based on the voltage sensed on the column.

Abstract: A binary CIM circuit enables all memory cells in a memory array to be effectively accessible simultaneously for computation using fixed pulse widths on the wordlines and equal capacitance on the bitlines. The fixed pulse widths and equal capacitance ensure that a minimum voltage drop in the bitline represents one least significant bit (LSB) so that the bitline voltage swing remains safely within the maximum allowable range. The binary CIM circuit maximizes the effective memory bandwidth of a memory array for a given maximum voltage range of bitline voltage.

Abstract: In one embodiment, a method comprises determining that a membrane potential of a first neuron of a first neuron core exceeds a threshold; determining a first plurality of synapse cores that each store at least one synapse weight associated with the first neuron; and sending a spike message to the determined first plurality of synapse cores.

Abstract: The present disclosure is directed to systems and methods of implementing a neural network using in-memory mathematical operations performed by pipelined SRAM architecture (PISA) circuitry disposed in on-chip processor memory circuitry. A high-level compiler may be provided to compile data representative of a multi-layer neural network model and one or more neural network data inputs from a first high-level programming language to an intermediate domain-specific language (DSL). A low-level compiler may be provided to compile the representative data from the intermediate DSL to multiple instruction sets in accordance with an instruction set architecture (ISA), such that each of the multiple instruction sets corresponds to a single respective layer of the multi-layer neural network model. Each of the multiple instruction sets may be assigned to a respective SRAM array of the PISA circuitry for in-memory execution.

Abstract: A binary CIM circuit enables all memory cells in a memory array to be effectively accessible simultaneously for computation using fixed pulse widths on the wordlines and equal capacitance on the bitlines. The fixed pulse widths and equal capacitance ensure that a minimum voltage drop in the bitline represents one least significant bit (LSB) so that the bitline voltage swing remains safely within the maximum allowable range. The binary CIM circuit maximizes the effective memory bandwidth of a memory array for a given maximum voltage range of bitline voltage.

Abstract: A loaded capacitance static random-access memory (C-SRAM) is provided. The C-SRAM is configured to prevent full bit line discharge during a functional reads even where the number of bit cells on the bit lines is small. The C-SRAM includes one or more loaded capacitance structures that may take any of a variety of physical configurations designed to provide additional capacitance to the bit lines. For instance, the loaded capacitance structures may take the form of a MIM capacitor in which a ferroelectric layer is formed from one or more high K materials. In addition, the loaded capacitance structures may be positioned in a variety of locations within the C-SRAM, including the back end of line.

Abstract: In one embodiment, an apparatus comprises a multiplier circuit to: identify a point multiply operation to be performed by the multiplier circuit, wherein the point multiply operation comprises point multiplication of a first plurality of operands; identify a point add operation associated with the point multiply operation, wherein the point add operation comprises point addition of a second plurality of operands, wherein the second plurality of operands comprises a first point and a second point, and wherein the first point and the second point are associated with a first coordinate system; convert the second point from the first coordinate system to a second coordinate system; perform the point add operation based on the first point associated with the first coordinate system and the second point associated with the second coordinate system; and perform the point multiply operation based on a result of the point add operation.

Abstract: Embodiments herein describe techniques for a semiconductor device including a memory cell. The memory cell includes a storage cell and a capacitor having a first electrode and a second electrode. The first electrode and the second electrode may be placed in a metal layer below a metal electrode coupled to one or more transistors of the storage cell. The storage cell is to store a digital value, where a voltage value of an output line of the storage cell is to be determined based on the digital value stored in the storage cell. The second electrode of the capacitor is coupled to the output line of the storage cell. The capacitor is to store a charge based on the voltage value of the output line of the storage cell. Other embodiments may be described and/or claimed.

Abstract: A DIMA semiconductor structure is disclosed. The DIMA semiconductor structure includes a frontend including a semiconductor substrate, a transistor switch of a memory cell coupled to the semiconductor substrate and a computation circuit on the periphery of the frontend coupled to the semiconductor substrate. Additionally, the DIMA includes a backend that includes an RRAM component of the memory cell that is coupled to the transistor switch.

Abstract: An embodiment of a semiconductor package apparatus may include technology to provide an entropy extractor including a physically unclonable function (PUF), a true random number generator (TRNG), an entropy source coupled to both the PUF and the TRNG, and a circuit to self-calibrate the entropy extractor. Other embodiments are disclosed and claimed.

Abstract: Embodiments herein describe techniques for a semiconductor device including a RRAM memory cell. The RRAM memory cell includes a substrate, a RRAM storage cell above the substrate, and a diode adjacent to the RRAM storage cell. The RRAM storage cell includes a first electrode located in a first metal layer above the substrate, a resistive switching material layer adjacent to the first electrode, and a second electrode adjacent to the resistive switching material layer. The second electrode is shared between the RRAM storage cell and the diode. The diode includes the second electrode shared with the RRAM storage cell, a semiconductor layer adjacent to the second electrode, and a third electrode located in a second metal layer above the substrate. Other embodiments may be described and/or claimed.

Abstract: Embodiments herein describe techniques for a semiconductor device including a RRAM memory cell. The RRAM memory cell includes a FinFET transistor and a RRAM storage cell. The FinFET transistor includes a fin structure on a substrate, where the fin structure includes a channel region, a source region, and a drain region. An epitaxial layer is around the source region or the drain region. A RRAM storage stack is wrapped around a surface of the epitaxial layer. The RRAM storage stack includes a resistive switching material layer in contact and wrapped around the surface of the epitaxial layer, and a contact electrode in contact and wrapped around a surface of the resistive switching material layer. The epitaxial layer, the resistive switching material layer, and the contact electrode form a RRAM storage cell. Other embodiments may be described and/or claimed.

Abstract: A compute near memory binary neural network accelerator with digital circuits that achieves energy efficiencies comparable to or surpassing a compute near memory binary neural network accelerator with analog circuits is provided. The compute near memory binary neural network accelerator with digital circuits is more process scalable, robust to process, voltage and temperature variations, and immune to circuit noise.

Abstract: In one embodiment, an apparatus comprises a multiplier circuit to: identify a plurality of partial products associated with a multiply operation; partition the plurality of partial products into a first set of partial products, a second set of partial products, and a third set of partial products; determine whether the multiply operation is associated with a square operation; upon a determination that the multiply operation is associated with the square operation, compute a result based on the first set of partial products and the third set of partial products; and upon a determination that the multiply operation is not associated with the square operation, compute the result based on the first set of partial products, the second set of partial products, and the third set of partial products.

Abstract: Techniques and mechanisms for providing data to be used in an in-memory computation at a memory device. In an embodiment a memory device comprises a first memory array and circuitry, coupled to the first memory array, to perform a data computation based on data stored at the first memory array. Prior to the computation, the first memory array receives the data from a second memory array of the memory device. The second memory array extends horizontally in parallel with, but is offset vertically from, the first memory array. In another embodiment, a single integrated circuit die includes both the first memory array and the second memory array.

Abstract: A compute near memory (CNM) convolution accelerator enables a convolutional neural network (CNN) to use dedicated acceleration to achieve efficient in-place convolution operations with less impact on memory and energy consumption. A 2D convolution operation is reformulated as 1D row-wise convolution. The 1D row-wise convolution enables the CNM convolution accelerator to process input activations row-by-row, while using the weights one-by-one. Lightweight access circuits provide the ability to stream both weights and input row as vectors to MAC units, which in turn enables modules of the CNM convolution accelerator to implement convolution for both [1×1] and chosen [n×n] sized filters.

Abstract: A memory circuit includes a number (X) of multiply-accumulate (MAC) circuits that are dynamically configurable. The MAC circuits can either compute an output based on computations of X elements of the input vector with the weight vector, or to compute the output based on computations of a single element of the input vector with the weight vector, with each element having a one bit or multibit length. A first memory can hold the input vector having a width of X elements and a second memory can store the weight vector. The MAC circuits include a MAC array on chip with the first memory.

Abstract: An apparatus is described. The apparatus includes a circuit to process a binary neural network. The circuit includes an array of processing cores, wherein, processing cores of the array of processing cores are to process different respective areas of a weight matrix of the binary neural network. The processing cores each include add circuitry to add only those weights of an i layer of the binary neural network that are to be effectively multiplied by a non zero nodal output of an i?1 layer of the binary neural network.

Abstract: An apparatus is described. The apparatus includes a long short term memory (LSTM) circuit having a multiply accumulate circuit (MAC). The MAC circuit has circuitry to rely on a stored product term rather than explicitly perform a multiplication operation to determine the product term if an accumulation of differences between consecutive, preceding input values has not reached a threshold.

Abstract: Computing devices and techniques for performing modular exponentiation for a data encryption process are described. In one embodiment, for example, an apparatus may include at least one memory logic for an encryption unit to perform encryption according to RSA encryption using a parallel reduction multiplier (PRM) MM process, at least a portion of the logic comprised in hardware coupled to the at least one memory and the at least one wireless transmitter, the logic to precompute a reduction coefficient, determine an operand product and a reduction product in parallel, the reduction product based on the reduction coefficient, and generate a MM result for the PRM MM process based on the operand product and the reduction product. Other embodiments are described and claimed.