Abstract: An approach provides a semiconductor structure for a first device with a first plurality of channels with a larger horizontal dimension than a vertical dimension of the first plurality of channels a second device comprising a second plurality of channels with a smaller horizontal dimension than the vertical dimension of the second plurality of channels. The first plurality of channels and the second plurality of channels have a same channel width in embodiments of the present invention. The first device is an n-type horizontal gate-all-around device and the second device is a p-type horizontal gate-all-around device.

Abstract: A memory device is provided. The memory device includes a ReRAM memory element, and a PCM memory element that is electrically connected in parallel with the ReRAM memory element.

Abstract: An approach for a nanosheet device with a single diffusion break is disclosed. The device comprises of active gate is formed above the BDI. At least the SDB is also formed over BDI with dielectric filled gate. The dielectric fill forms an indentation into the remaining nanosheets, under the spacer region, or between the inner spacers, in the SDB region. The method of creating the device comprises of, forming a gate cut opening between two ends of a dummy gate of one or more gates; forming a first sacrificial material on the gate cut opening; creating a single diffusion break; removing the dummy gate and oxide layer; removing, selectively a second sacrificial material; trimming, selectively stack of nanosheets; and forming dielectric in the gate cut opening and the single diffusion break.

Abstract: Embodiments of the present invention include a memory cell that has a vertically-oriented fin. The memory cell may also include a resistive memory device located on a first lateral side of the fin. The resistive memory device may include a bottom electrode, a top electrode, and a resistive element between the bottom electrode and the top electrode. The memory cell may also include a vertical field-effect transistor having a metal gate and a gate dielectric contacting a second lateral side of the fin opposite the first lateral side.

Abstract: A ring-shaped heater, system, and method to gradually change the conductance of the phase change memory through a concentric ring-shaped heater. The system may include a phase change memory. The phase change memory may include a bottom electrode. The phase change memory may also include a ring-shaped heater patterned on top of the bottom electrode, the ring-shaped heater including: a plurality of concentric conductive heating layers, and a plurality of insulator spacers, where each insulator spacer separates each conductive heating layer. The phase change memory may also include a phase change material proximately connected to the ring-shaped heater. The phase change memory may also include a top electrode proximately connected to the phase change material.

Abstract: A phase change memory, system, and method for gradually changing the conductance and resistance of the phase change memory while preventing resistance drift. The phase change memory may include a phase change material. The phase change memory may also include a bottom electrode. The phase change memory may also include a heater core proximately connected to the bottom electrode. The phase change memory may also include a set of conductive rings surrounding the heater core, where the set of conductive rings comprises one or more conductive rings, and where the set of conductive rings are proximately connected to the phase change material. The phase change memory may also include a set of spacers, where a spacer, from the set of spacers, separates a portion of a conductive ring, from the set of conductive rings, from the heater core.

Abstract: Embodiments of the present invention include a memory cell that has a vertically-oriented fin. The memory cell may also include a resistive memory device located on a first lateral side of the fin. The resistive memory device may include a bottom electrode, a top electrode, and a resistive element between the bottom electrode and the top electrode. The memory cell may also include a vertical field-effect transistor having a metal gate and a gate dielectric contacting a second lateral side of the fin opposite the first lateral side.

Abstract: A phase change memory (PCM) cell comprising a substrate a first electrode located on the substrate. A phase change material layer located adjacent to the first electrode, wherein a first side of the phase change material layer is in direct contact with the first electrode. A second electrode located adjacent to phase change material layer, wherein the second electrode is in direct contact with a second side of the phase change material layer, wherein the first side and the second side are different sides of the phase change material layer. An airgap is located directly above the phase change material layer, wherein the airgap provides space for the phase change material to expand or restrict.

Abstract: A semiconductor structure for a vertical phase change memory cell that includes a bottom electrode on a portion of a semiconductor substrate and a pair of vertical phase change bridge elements that are each on a portion of the bottom electrode. The semiconductor structure for the vertical phase change memory cell includes a dielectric material separating the pair of vertical phase change bridge elements and a top electrode over the pair of vertical phase change bridge elements.

Abstract: A method of fabricating a static random-access memory (SRAM) device includes forming a sacrificial material and replacing the sacrificial material with a metal to form a cross-couple contact on a metal gate stack. A portion of the metal gate stack directly contacts each of a sidewall and an endwall of the cross-couple contact.

Abstract: A method of fabricating a static random-access memory (SRAM) device includes forming a sacrificial material and replacing the sacrificial material with a metal to form a cross-couple contact on a metal gate stack. A portion of the metal gate stack directly contacts each of a sidewall and an endwall of the cross-couple contact.

Abstract: A non-volatile memory structure, and methods of manufacture, which may include a first memory element and a second memory element between a first terminal and a second terminal. The first memory element and the second memory element may be in parallel with each other between the first and second terminal. This may enable the hybrid non-volatile memory structure to store values as a combination of the conductance for each memory element, thereby enabling better tuning of set and reset conductance parameters.

Abstract: A semiconductor device includes a base structure including a lower level via and a lower level dielectric layer, a conductive pillar including an upper level line and an upper level via disposed on the lower level via, and a protective structure disposed between the lower level via and the upper level line. The protective structure includes a material having an etch rate less than or equal to that of the lower level via.

Abstract: Embodiments of the present invention are directed to methods and resulting structures for nanosheet devices having asymmetric gate stacks. In a non-limiting embodiment of the invention, a nanosheet stack is formed over a substrate. The nanosheet stack includes alternating semiconductor layers and sacrificial layers. A sacrificial liner is formed over the nanosheet stack and a dielectric gate structure is formed over the nanosheet stack and the sacrificial liner. A first inner spacer is formed on a sidewall of the sacrificial layers. A gate is formed over channel regions of the nanosheet stack. The gate includes a conductive bridge that extends over the substrate in a direction orthogonal to the nanosheet stack. A second inner spacer is formed on a sidewall of the gate. The first inner spacer is formed prior to the gate stack, while the second inner spacer is formed after, and consequently, the gate stack is asymmetrical.

Abstract: Embodiments of the present invention are directed to methods and resulting structures for nanosheet devices having asymmetric gate stacks. In a non-limiting embodiment of the invention, a nanosheet stack is formed over a substrate. The nanosheet stack includes alternating semiconductor layers and sacrificial layers. A sacrificial liner is formed over the nanosheet stack and a dielectric gate structure is formed over the nanosheet stack and the sacrificial liner. A first inner spacer is formed on a sidewall of the sacrificial layers. A gate is formed over channel regions of the nanosheet stack. The gate includes a conductive bridge that extends over the substrate in a direction orthogonal to the nanosheet stack. A second inner spacer is formed on a sidewall of the gate. The first inner spacer is formed prior to the gate stack, while the second inner spacer is formed after, and consequently, the gate stack is asymmetrical.

Abstract: A method for using a distributed memory device in a memory augmented neural network system includes receiving, by a controller, an input query to access data stored in the distributed memory device, the distributed memory device comprising a plurality of memory banks. The method further includes determining, by the controller, a memory bank selector that identifies a memory bank from the distributed memory device for memory access, wherein the memory bank selector is determined based on a type of workload associated with the input query. The method further includes computing, by the controller and by using content based access, a memory address in the identified memory bank. The method further includes generating, by the controller, an output in response to the input query by accessing the memory address.

Abstract: Embodiments of the present invention are directed to methods and resulting structures for nanosheet devices having asymmetric gate stacks. In a non-limiting embodiment of the invention, a nanosheet stack is formed over a substrate. The nanosheet stack includes alternating semiconductor layers and sacrificial layers. A sacrificial liner is formed over the nanosheet stack and a dielectric gate structure is formed over the nanosheet stack and the sacrificial liner. A first inner spacer is formed on a sidewall of the sacrificial layers. A gate is formed over channel regions of the nanosheet stack. The gate includes a conductive bridge that extends over the substrate in a direction orthogonal to the nanosheet stack. A second inner spacer is formed on a sidewall of the gate. The first inner spacer is formed prior to the gate stack, while the second inner spacer is formed after, and consequently, the gate stack is asymmetrical.

Abstract: In a deep neural network (DNN), weights are defined that represent a strength of connections between different neurons of the DNN and activations are defined that represent an output produced by a neuron after passing through an activation function of receiving an input and producing an output based on some threshold value. The weight traffic associated with a hybrid memory therefore is distinguished from the activation traffic to the hybrid memory, and one or more data structures may be dynamically allocated in the hybrid memory according to the weights and activations of the or more data structures in the DNN. The hybrid memory includes at least a first memory and a second memory that differ according to write endurance attributes.

Abstract: In a deep neural network (DNN), weights are defined that represent a strength of connections between different neurons of the DNN and activations are defined that represent an output produced by a neuron after passing through an activation function of receiving an input and producing an output based on some threshold value. The weight traffic associated with a hybrid memory therefore is distinguished from the activation traffic to the hybrid memory, and one or more data structures may be dynamically allocated in the hybrid memory according to the weights and activations of the one or more data structures in the DNN. The hybrid memory includes at least a first memory and a second memory that differ according to write endurance attributes.

Abstract: A method for using a distributed memory device in a memory augmented neural network system includes receiving, by a controller, an input query to access data stored in the distributed memory device, the distributed memory device comprising a plurality of memory banks. The method further includes determining, by the controller, a memory bank selector that identifies a memory bank from the distributed memory device for memory access, wherein the memory bank selector is determined based on a type of workload associated with the input query. The method further includes computing, by the controller and by using content based access, a memory address in the identified memory bank. The method further includes generating, by the controller, an output in response to the input query by accessing the memory address.