A semiconductor memory cell and arrays of memory cells are provided In at least one embodiment, a memory cell includes a substrate having a top surface, the substrate having a first conductivity type selected from a p-type conductivity type and an n-type conductivity type; a first region having a second conductivity type selected from the p-type and n-type conductivity types, the second conductivity type being different from the first conductivity type, the first region being formed in the substrate and exposed at the top surface; a second region having the second conductivity type, the second region being formed in the substrate, spaced apart from the first region and exposed at the top surface; a buried layer in the substrate below the first and second regions, spaced apart from the first and second regions and having the second conductivity type; a body region formed between the first and second regions and the buried layer, the body region having the first conductivity type; a gate positioned between the first and second regions and above the top surface; and a nonvolatile memory configured to store data upon transfer from the body region.

Methods, systems, and devices for dirty write on power off are described. In an example, the described techniques may include writing memory cells of a device according to one or more parameters (e.g., reset current amplitude), where each memory cell is associated with a storage element storing a value based on a material property associated with the storage element. Additionally, the described techniques may include identifying, after writing the memory cells, an indication of power down for the device and refreshing, before the power down of the device, a portion of the memory cells based on identifying the indication of the power down for the device. In some cases, refreshing includes modifying at least one of the one or more parameters for a write operation for the portion of the memory cells.

An improved electronic mixed mode multiplier and accumulate circuit for artificial intelligence and computing system applications that perform vector-vector, vector-matrix and other multiply-accumulate computations. The circuit is provided is a high resolution, high linearity, low area, low power multiply—accumulate (MAC) unit to interface with a memory device for storing computation output results. The MAC unit uses a less number of current carrying elements resulting in much lower integrated circuit area, and provides a tight matching between the current elements thus preserving inherent linearity requirements due to current mode operation. Further the MAC performs current scaling using switches and current division where the current switches occupy minimum size transistors requiring a small area to implement that renders it compatible with MRAM such as a magnetic tunnel junction device. The MAC is hierarchically extended for increased number of bits to provide a delay implementation using orthogonal vector and current addition.

Techniques for burst memory write operations are disclosed. In the illustrative embodiment, a memory die is limited in how quickly it can perform memory write operations that it receives from a microcontroller due to thermal constraints. The memory die can mitigate the need for the microcontroller to perform a costly rank switch to send an operation to another die by buffering memory write operations. The microcontroller can then send several consecutive memory write operations to a first memory die before switching to a second memory die. The first memory die can then perform the memory write operations while the microcontroller has moved on to other memory operations.

A read technique for both SLC (single level cell) and MLC (multi-level cell) cross-point memory can mitigate drift-related errors with minimal or no drift tracking. In one example, a read at a higher magnitude voltage is applied first, which causes the drift for cells in a lower threshold voltage state to be reset. In one example, the read at the first voltage can be a full float read to minimize disturb. A second read can then be performed at a lower voltage without the need to adjust the read voltage due to drift.

Methods, systems, and devices for programming techniques for polarity-based memory cells are described. A method may include writing memory cells to an intermediate state based on receiving a write command. Writing the intermediate state may include applying a first pulse having a first polarity to the memory cell. The method may include isolating a first access line coupled with the memory cell from a voltage source based on applying the first pulse. The method may also include applying a second pulse to a second access line coupled with the memory cell based on isolating the first access line.

A method for training an artificial neural network circuit is provided. The artificial neural network circuit includes a crossbar circuit that has a plurality of input bars, a plurality of output bars crossing the plurality of input bars, and memristors each of which includes a variable conductance element provided at corresponding one of intersections of the input bars and the output bars.

An output, representing synaptic weights of a neural network can be received from first memory elements. The output can be compared to a known correct output. A random number can be generated with a tuned bias via second memory elements. The weights can be updated based on the random number and a difference between the output and the known correct output.

A method is disclosed that includes causing a first set of a plurality of voltage pulses to be applied to memory cells of a memory device, a voltage pulse of the first set of the voltage pulses placing the memory cells of the memory device at a voltage level associated with a defined voltage state. The method also includes determining a set of bit error rates associated with the memory cells of the memory device in view of a data mapping pattern for the memory cells of the memory device, wherein the data mapping pattern assigns a voltage level associated with a reset state to at least a portion of the memory cells of the memory device. The method further includes determining whether to apply one or more second sets of the voltage pulses to the memory cells of the memory device in view of a comparison between the set of bit error rates for the memory cells and a previously measured set of bit error rates for the memory cells.

Methods, apparatuses, and systems for in-or near-memory processing are described. Bits of a first number may be stored on a number of memory elements, wherein each memory element of the number of memory elements intersects a bit line and a word line of a number of word lines. A number of signals corresponding to bits of a second number may be driven on the number of word lines to generate a number of output signals. A value equal to a product of the first number and the second number may be generated based on the number of output signals.