Abstract: Techniques are provided to compensate for neighbor word line interference when programming memory cells connected to a selected word line WLn. Before programming, the assigned data states of WLn and WLn+1 are compared and corresponding compensation data is generated. The compensation data may be stored in latches of sense circuits to modify the verify tests which occur during programming. The compensation can involve adjusting the bit line voltage, word line voltage, sense node discharge period and/or trip voltage. During a verify test, the compensation data can cause a WLn memory cell to complete programming when its threshold voltage is lower than would be the case with no compensation. When the WLn+1 memory cells are subsequently programmed, an upshift in the threshold voltage of the WLn memory cell offsets the compensation.

Abstract: Techniques are provided for pre-charging NAND strings during a programming operation. The NAND strings are in a block that is divided into vertical sub-blocks. During a pre-charge phase of a programming operation, an overdrive voltage is applied to some memory cells and a bypass voltage is applied to other memory cells. The overdrive voltage allows the channel of an unselected NAND string to adequately charge during the pre-charge phase. Adequate charging of the channel helps the channel voltage to boost to a sufficient level to inhibit programming of an unselected memory cell during a program phase. Thus, program disturb is prevented, or at least reduced. The technique allows, for example, programming of memory cells in a middle vertical sub-block without causing program disturb of memory cells that are not to receive programming.

Abstract: A method of concurrently programming a memory. Various methods include: applying a non-negative voltage on a first bit line coupled to a first memory cell; applying a negative voltage on a second bit line coupled to a second memory cell, where the negative voltage is generated using triple-well technology; then applying a programming pulse to the first and second memory cells concurrently; and in response, programming the first and second memory cells to different states. The methods also include applying a quick pass write operation to the first and second memory cells, by: applying a quick pass write voltage to the first bit line coupled to the first memory cell, where the quick pass write voltage is higher than the non-negative voltage; applying a negative quick pass write voltage to the second bit line coupled to the first memory cell, where the negative quick pass write voltage is generated using triple-well technology.

Abstract: A method for reducing firmware size and increasing firmware performance. Core timing control conditions used by a die controller are converted into production ready core timing control conditions, from which firmware instructions are then generated. The production ready core timing control conditions comprise a plurality of fixed core timing control conditions. The firmware instructions are modified to determine core timing control condition values for fixed core timing control conditions before implementing storage operations, to store the core timing control condition values in global condition registers, and to modify references to fixed core timing control conditions to access the values in those global condition registers. Finally, the modified firmware instructions are stored on the die controller, which comprises a microcontroller configured to execute them.

Abstract: An illustrative embodiment disclosed herein is an apparatus including a non-volatile memory cell and multi-bit input circuitry that simultaneously receives a plurality of bits, receives a supply voltage, converts the plurality of bits and the supply voltage into a multiply voltage, and applies the multiply voltage to the non-volatile memory cell. The non-volatile memory cell may pass a memory cell current in response to the multiply voltage. A magnitude of the multiply voltage may represent a multiplier. The memory cell current may represent a product of the multiplier and a multiplicand stored in the non-volatile memory cell.

Abstract: Apparatuses, systems, and methods are disclosed for concurrently programming non-volatile storage cells, such as those of an SLC NAND array. The non-volatile storage cells may be arranged into a first block comprising a first string of storage cells that intersects with a first word line at a first storage cell, a second block comprising a second string of storage cells that intersects with a second word line at a second storage cell, a bit line electrically connectable to the first string and the second string, and controller configured to apply a programming pulse, at an elevated voltage, to the first word line and second word line to concurrently program the first and second storage cells.

Abstract: A memory array is provided that includes a first memory level having a plurality of memory cells each including a corresponding magnetic memory element coupled in series with a corresponding selector element, and a plurality of vias, each of the vias coupled in series with a corresponding one of the memory cells.

Abstract: A memory apparatus and method of operation is provided. The apparatus includes memory cells coupled to a control circuit. The control circuit is configured to perform a first programming stage including iteratively programming each of the memory cells to first program states and verifying that the memory cells have a threshold voltage above one of a plurality of first verify voltages corresponding to the first program states. The first programming stage ends before all of the memory cells are verified thereby leaving a fraction of the memory cells below the one of the plurality of first verify voltages. The control circuit also performs a second programming stage including iteratively programming each of the memory cells to second program states and verifying that at least a predetermined number of the memory cells have the threshold voltage above one of a plurality of second verify voltages corresponding to the second program states.

Abstract: Method(s) and structure(s) for a two-page read operation are described and provide a multiple page read. The two page read operation provides for reading two pages with in a block without reducing the control gates to a low voltage level. The two page read can read the first page using an incrementing voltage level at discrete steps and starting the second page read at the high state for the control gates from the first page read. The second page read then decrements the control gate voltages level through the steps. This should reduce energy consumption. The two-page read operation will also reduce the time as the time period to reset the control gates to a low state are not required in between the page read operations.

Abstract: A circuit comprising a non-volatile memory array, an Input/Output (IO) circuit, a decoder circuit, a control circuit, and a read/write circuit. The non-volatile memory array couples to an address decoder that identifies a location within the non-volatile memory array for a storage command. The IO circuit couples to a decoder circuit through a control bus. The decoder circuit decodes a command address and storage command from a fixed length command sequence received by the IO circuit over the data bus. The decoder circuit may include a serial-in parallel out (SIPO) circuit for decoding and parallel operation. The control circuit couples to the IO and decoder circuits and generates control signals to execute decoded storage commands. The read/write circuit couples to the non-volatile memory array and the control circuit. The read/write circuit transfers data between the non-volatile memory array and the IO circuit in response to the storage commands.

Abstract: Apparatuses and techniques are described for forming of selectors in a memory device such as a crosspoint memory array. A threshold switching selector is in series with a resistance-switching memory cell in a storage node. Prior to a first switching operation in the array, a stimulus is applied to the storage node to transform the selectors from an initial state having an initial threshold voltage to an operating state having a lower, operating threshold voltage. The stimulus can include a signal having a voltage which does not exceed the operating threshold voltage. To limit peak current consumption, the stimulus can be applied to different subsets of the array, one subset at a time.

Abstract: A sense amplifier for a memory circuit is presented that can sense a selected memory cell in either a first sensing mode, in which current from the selected memory cell flows from the memory cells into the sense amplifier, or a second sensing mode, in which current is discharged from the sense amplifier through the selected memory cell. In the first sensing mode, current from a selected memory cell is conducted through cascaded PMOS transistors to charge a sensing node, with the resultant voltage level on the sensing node used to determine the result of the sensing operation.

Abstract: Techniques are presented for performing in-memory matrix multiplication operations for binary input, binary weight valued convolution neural network (CNN) inferencing. The weights of a filter are stored in pairs of memory cells of a storage class memory device, such as a ReRAM or phase change memory based devices. To reduce current consumption, the binary valued filters are transformed into ternary valued filters by taking sums and differences of binary valued filter pairs. The zero valued weights of the transformed filters are stored as a pair of high resistance state memory cells, reducing current consumption during convolution. The results of the in-memory multiplications are pair-wise combined to compensate for the filter transformations. To compensate for zero valued weights, a zero weight register stores the number of zero weights along each bit line and is used to initialize counter values for accumulating the multiplication operations.

Abstract: In a non-volatile memory circuit, performance is improved by converting data between a serial format, for transfer on and off of the memory circuit, and a parallel format, for transfer to and from the memory latches used for read and writing data into the memory array of the memory circuit. The memory array is split into M+N divisions, but transferred with a degree of parallelism of M, allowing M words of data to be transferred in parallel at a fixed transfer rate while allowing for up to N bad columns in a transfer. In the write path, a column skipping mechanism is used when converting words of write data into a parallel format. In the read path, a set of (M+N) to 1 multiplexers is used to align the word of read data so that read data can be transferred at a fixed rate and without any added latency.

Abstract: A sense amplifier for a memory circuit is presented that can reduce sensing times by introduction of a local reference generator. The sense amplifier includes two capacitors that are pre-charged prior to a sensing operation. A first of the capacitors is connected so that it can discharge through a selected memory cell at a rate dependent on the conductivity of the selected memory cell. After a sensing interval in which the first capacitor can discharge through the selected memory cell, the voltage level on the first capacitor is compared with the voltage level on the second capacitor to determine the result of the sensing operation.

Abstract: Apparatuses and techniques are described for periodically refreshing word line voltages in a group of blocks in a memory device. In one aspect, each group of blocks stores the same number of bits per cell. For example, one group of blocks can be reserved for single level cell (SLC) data and another group of blocks can be reserved for multi-level cell (MLC) data. A common refresh voltage signal can be applied to the blocks in a group, where the voltage signal is optimized based on the number of bits per cell stored by the memory cells of the group. For an SLC block, the refresh voltage signal can decrease a floating voltage of the word lines. For an MLC block, the refresh voltage signal can increase a floating voltage of the word lines.

Abstract: A 3D NAND memory device is provided in which control is performed by two microcontroller units (MCU). During manufacture of the memory device, bug fixes required for the controller may be addressed using a software solution by which an instruction requiring correction in one of the two MCUs is replaced with a corrected instruction stored in a RAM.

Abstract: Apparatuses and techniques are described for providing a positive voltage source and a negative voltage source in a circuit. The positive voltage source and the negative voltage source have a common ground node. The positive voltage source can be provided using a current mirror in which a current in a first path is copied to provide a current in a second path. The currents of the first and second paths are sunk at the common ground node. The negative voltage source can be provided using a current mirror in which a current in a third path is copied to provide a current in a fourth path, where the current of the fourth path is sourced at the common ground node.

Abstract: Apparatuses and techniques are described for reading crosspoint arrays of memory cells with high bandwidth and a relatively small page buffer. Multiple crosspoint arrays (XPAs) are read in parallel, with one memory cell per XPA being read, in a bank of XPAs. To reduce the read time, a row can be selected for the XPAs, after which memory cells in different columns are read, one column at a time, while the same row is selected. This avoids the need to transmit commands and a row address for re-selecting the row in each successive read operation. The XPAs may be ungrouped, or one XPA may be accessible at a time in a group. In one option, the XPAs are arranged in sets, either individually or in groups, and one set is accessible at a time.

Abstract: A method of operating a storage device, including; performing, by a non-volatile memory, an erase operation on a block of memory in the non-volatile memory, where the non-volatile memory is coupled to a controller; receiving, by the non-volatile memory, a host-transaction within a first time period, where, the non-volatile memory is coupled to a host device; and suspending, by the non-volatile memory, an erase operation in response to receiving the host-transaction by: determining the erase operation has completed a charge phase; and suspending the erase operation during a pulse phase of the erase operation. The method additionally includes the non-volatile memory maintaining a loop counter and a pulse counter, where: the loop counter increments in response to completion of an erase loop, and the pulse counter increments in response to completion of an erase pulse, where the erase pulse is applied during a pulse phase of the erase operation.