Abstract:
A semiconductor device for a one-time programmable (OTP) memory according to some examples of the disclosure includes a gate, a dielectric region below the gate, a source terminal below the dielectric region and offset to one side, a drain terminal below the dielectric region and offset to an opposite side from the source terminal, a drain side charge trap in the dielectric region capable of programming the semiconductor device, and a source side charge trap in the dielectric region opposite the drain side charge trap and capable of programming the semiconductor device.
Abstract:
A complementary metal oxide semiconductor (CMOS) static random access memory (SRAM) cell. A CMOS SRAM cell in accordance with an aspect of the present disclosure includes a bit line and a word line. Such a CMOS SRAM memory cell further includes a CMOS memory cell having at least a first p-channel device comprising a first channel material that differs from a substrate material of the CMOS memory cell, the first channel material having an intrinsic channel mobility greater than the intrinsic channel mobility of the substrate material, the first p-channel device coupling the CMOS memory cell to the bit line and the word line.
Abstract:
Features are fabricated on a semiconductor chip. The features are smaller than the threshold of the lithography used to create the chip. A method includes patterning a first portion of a feature (such as a local interconnect) and a second portion of the feature to be separated by a predetermined distance, such as a line tip to tip space or a line space. The method further includes patterning the first portion with a cut mask to form a first sub-portion (e.g., a contact) and a second sub-portion. A dimension of the first sub-portion is less than a dimension of a second predetermined distance, which may be a line length resolution of a lithographic process having a specified width resolution. A feature of a semiconductor device includes a first portion and a second portion having a dimension less than a lithographic resolution of the first portion.
Abstract:
A complementary metal oxide semiconductor (CMOS) static random access memory (SRAM) cell. A CMOS SRAM cell in accordance with an aspect of the present disclosure includes a bit line and a word line. Such a CMOS SRAM memory cell further includes a CMOS memory cell having at least a first p-channel device comprising a first channel material that differs from a substrate material of the CMOS memory cell, the first channel material having an intrinsic channel mobility greater than the intrinsic channel mobility of the substrate material, the first p-channel device coupling the CMOS memory cell to the bit line and the word line.
Abstract:
Methods and apparatus for performing machine learning tasks, and in particular, to a neural-network-processing architecture and circuits for improved handling of partial accumulation results in weight-stationary operations, such as operations occurring in compute-in-memory (CIM) processing elements (PEs). One example PE circuit for machine learning generally includes an accumulator circuit, a flip-flop array having an input coupled to an output of the accumulator circuit, a write register, and a first multiplexer having a first input coupled to an output of the write register, having a second input coupled to an output of the flip-flop array, and having an output coupled to a first input of the first accumulator circuit.
Abstract:
A stacked system-on-chip (SoC) is described. The stacked SoC includes a first memory die comprising a dynamic random-access memory (DRAM). The stacked SoC also includes a compute logic die. The compute logic die comprises a static random-access memory (SRAM) having a first SRAM partition and a second SRAM partition. The first memory die is stacked on the compute logic die. The compute logic die includes a memory controller. The memory controller is coupled between the first SRAM partition and the second SRAM partition. Additionally, the memory controller is coupled to a DRAM bus of the first memory die.
Abstract:
Certain aspects provide methods and apparatus for binary computation. An example circuit for such computation generally includes a memory cell having at least one of a bit-line or a complementary bit-line; a computation circuit coupled to a computation input node of the circuit and the bit-line or the complementary bit-line; and an adder coupled to the computation circuit, wherein the computation circuit comprises a first n-type metal-oxide-semiconductor (NMOS) transistor coupled to the memory cell, and a first p-type metal-oxide-semiconductor (PMOS) transistor coupled to the memory cell, drains of the first NMOS and PMOS transistors being coupled to the adder, wherein a source of the first PMOS transistor is coupled to a reference potential node, and wherein a source of the first NMOS transistor is coupled to the computation input node.
Abstract:
An apparatus includes first and second compute-in-memory (CIM) arrays. The first CIM array is configured to store weights corresponding to a filter tensor, to receive a first set of activations corresponding to a first receptive field of an input, and to process the first set of activations with the weights to generate a corresponding first tensor of output values. The second CIM array is configured to store a first copy of the weights corresponding to the filter tensor and to receive a second set of activations corresponding to a second receptive field of the input. The second CIM array is also configured to process the second set of activations with the first copy of the weights to generate a corresponding second tensor of output values. The first and second compute-in-memory arrays are configured to process the first and second receptive fields in parallel.
Abstract:
Systems and methods relate to a seven transistor static random-access memory (7T SRAM) bit cell which includes a first inverter having a first pull-up transistor, a first pull-down transistor, and a first storage node, and a second inverter having a second pull-up transistor, a second pull-down transistor, and a second storage node. The second storage node is coupled to gates of the first pull-up transistor and the first pull-down transistor. A transmission gate is configured to selectively couple the first storage node to gates of the second pull-up transistor and the second pull-down transistor during a write operation, a standby mode, and a hold mode, and selectively decouple the first storage node from gates of the first pull-up transistor and a first pull-down transistor during a read operation. The 7T SRAM bit cell can be read or written through an access transistor coupled to the first storage node.
Abstract:
Methods and apparatus directed toward a high density static random access memory (SRAM) array having advanced metal patterning are provided. In an example, provided is a method for fabricating an SRAM. The method includes forming, using a self-aligning double patterning (SADP) technique, a plurality of substantially parallel first metal lines oriented in a first direction in a first layer. The method also includes etching the substantially parallel first metal lines, using a cut mask, in a second direction substantially perpendicular to the first direction, to separate the substantially parallel first metal lines into a plurality of islands having first respective sides that are aligned in the first direction and second respective sides that are aligned the second direction. The method also includes forming, in a second layer, a plurality of second metal lines oriented in the first direction.