According to one embodiment, a memory system includes: a first package including a first memory chip configured to store data, and a first chip containing a first circuit configured to control an On Die Termination (ODT) operation based on a first signal which is a control signal for reading of data stored in the first memory chip; a second package including a second memory chip configured to store data, and a second chip containing a second circuit configured to control the ODT operation based on the first signal, the first signal also being a control signal for reading of data stored in the second memory chip; and a controller configured to transmit the first signal to the first chip and the second chip.

A semiconductor memory device includes a memory core that performs reading and writing of data, data delivery and training blocks that are connected between first pads and the memory core, and at least one data delivery, clock generation and training block that is connected between at least one second pad and the memory core. In a first training operation, the data delivery and training blocks output first training data, received through the first pads, through the first pads as second training data. In a second training operation, at least one of the data delivery and training blocks outputs third training data, received through the at least one second pad, through at least one of the first pads as fourth training data. The second training data and the fourth training data are output in synchronization with read data strobe signals output through the at least one second pad.

Methods of performing mixed-signal/analog multiply-accumulate (MAC) operations used for matrix multiplication in fully connected artificial neural networks in integrated circuits (IC) are described in this disclosure having traits such as: (1) inherently fast and efficient for approximate computing due to current-mode signal processing where summation is performed by simply coupling wires, (2) free from noisy and power hungry clocks with asynchronous fully-connected operations, (3) saving on silicon area and power consumption for requiring neither any data-converters nor any memory for intermediate activation signals, (4) reduced dynamic power consumption due to Compute-In-Memory operations, (5) avoiding over-flow conditions along key signals paths and lowering power consumption by training MACs in neural networks in such a manner that the population and or combinations of multi-quadrant activation signals and multi-quadrant weight signals follow a programmable statistical distribution profile, (6) programmable current consumption versus degree of precision/approximate computing, (7) suitable for ‘always-on’ operations and capable of ‘self power-off’, (8) inherently simple arrangement for non-linear activation operations such as Rectified Linear Unit, ReLu, and (9) manufacturable on main-stream, low cost, and lagging edge standard digital CMOS process requiring neither any resistors nor any capacitors.

System on chips, memory circuits, and method for data access, the memory circuits including a memory cell array and an input/output (I/O) connection interface coupled to the memory cell array, wherein the I/O connection interface is configured for coupling to an external signal line to directly receive a transistor-level operation signal from an external memory controller for accessing data in the memory cell array.

A semiconductor package includes: a package board including a plurality of connection pads; a semiconductor chip including a first surface and a plurality of bonding pads, wherein the first surface of the semiconductor chip contacts a first surface of the package board, and wherein the plurality of bonding pads are respectively connected to the plurality of connection pads; and a thermal fuse circuit connected between a sensing connection pad of the plurality of connection pads and a sensing bonding pad of the plurality of bonding pads, and configured to open between the sensing connection pad and the sensing bonding pad when an internal temperature of the thermal fuse circuit is greater than or equal to a cutoff temperature of the thermal fuse circuit.

A lateral transfer path within an adjustable-width signaling interface of an integrated circuit component is formed by a chain of logic segments that may be intercoupled in different groups to effect the lateral data transfer required in different interface width configurations, avoiding the need for a dedicated transfer path per width configuration and thereby substantially reducing number of interconnects (and thus the area) required to implement the lateral transfer structure.

Methods, systems, and devices for stacked memory dice and combined access operations are described. A device may include multiple memory dice. One die may be configured as a master, and another may be configured as a slave. The master may communicate with a host device. A slave may be coupled with the master but not the host device. The device may include a first die (e.g., master) and a second die (e.g., slave). The first die may be coupled with a host device and configured to output a set of data in response to a read command. The first die may supply a first subset of the data and obtain a second subset of the data from the second die. In some cases, the first die may select, based on a data rate, a modulation scheme (e.g., PAM4, NRZ, etc.) and output the data using the selected modulation scheme.

A semiconductor package is provided. The semiconductor package includes: a substrate; a first buffer chip and a second buffer chip located on an upper part of the substrate; a plurality of nonvolatile memory chips located on the upper part of the substrate and including a first nonvolatile memory chip and a second nonvolatile memory chip, the first nonvolatile memory chip being electrically connected to the first buffer chip, and the second nonvolatile memory chip being electrically connected to the second buffer chip; a plurality of external connection terminals connected to a lower part of the substrate; and a rewiring pattern located inside the substrate. The rewiring pattern is configured to diverge an external electric signal received through one of the plurality of external connection terminals into first and second signals, transmit the first signal to the first buffer chip, and transmit the second signal to the second buffer chip.

Disclosed herein is a method for designing a semiconductor device, the method including: assigning a plurality of wiring tracks including first and second tracks; connecting a first data I/O circuit to a first data node of a first circuit by a first signal bus arranged on the first wiring track; connecting a second data I/O circuit to a second data node of the first circuit by a second signal bus arranged on the second wiring track when a first design mode is selected; and connecting the first data I/O circuit to a second circuit by a second signal bus arranged on the second wiring track when a second design mode is selected.

Methods, systems, and devices for on-die termination configuration for a memory device are described. In some examples, a memory device may determine a connection option from a set of connections options for which an ODT pin of the memory device is configured. Each connection option may correspond to a termination configuration for a different pin, such as a command and address (CA) pin, a clock (CK) pin, or a chip select (CS). Based on the determined connection option, the memory device may identify a respective termination option for each of the different pins, such as a first termination option for the CA pin, a second termination option for the CK pin, and a third termination option for the CS pin, and configure each of the different pins according to the respective termination option for that pin.