Abstract: A stacked memory device includes memory dies over a base die. The base die includes separate memory channels to the different dies and external channels that allow an external processor access to the memory channels. The base die allows the external processor to access multiple memory channels using more than one external channel. The base die also allows the external processor to communicate through the memory device via the external channels, bypassing the memory channels internal to the device. This bypass functionality allows the external processor to connect to additional stacked memory devices.

Abstract: A memory system includes a link having at least one signal line and a controller. The controller includes at least one transmitter coupled to the link to transmit first data, and a first error protection generator coupled to the transmitter. The first error protection generator dynamically adds an error detection code to at least a portion of the first data. At least one receiver is coupled to the link to receive second data. A first error detection logic determines if the second data received by the controller contains at least one error and, if an error is detected, asserts a first error condition. The system includes a memory device having at least one memory device transmitter coupled to the link to transmit the second data. A second error protection generator coupled to the memory device transmitter dynamically adds an error detection code to at least a portion of the second data.

Abstract: A system that includes a non-volatile memory subsystem having non-volatile memory. The system also includes a plurality of memory modules that are separate from the non-volatile memory subsystem. Each memory module can include a plurality of random access memory packages where each first random access memory package includes a primary data port and a backup data port. Each memory module can include a storage interface circuit coupled to the backup data ports of the random access memory packages. The storage interface circuit offloads data from the memory module in the event of a power loss by receiving data from the backup data ports of the random access memory packages and transmitting the data to the non-volatile memory subsystem.

Abstract: A binary receiver combines a fast amplifier with a relatively slow amplifier for noise rejection. Both the fast and slow amplifiers employ hysteresis. The fast amplifier has relatively lower hysteresis, meaning that its sensitivity is a less effected by prior data values but more susceptible to glitch-induced errors. Conversely, the slow amplifier has relatively higher hysteresis and rejects glitches but introduces undesirable signal-propagation delays. A state machine taking input from both amplifiers allows the receiver to filter glitches without incurring a significant data-propagation delay.

Abstract: A PAM-4 DFE receives an input signal distorted by inter-symbol interference (ISI) and expressing a series of symbols each representing one of four pulse amplitudes to convey two binary bits of data per symbol. High-order circuitry resolves the most-significant bit (MSB) of each two-bit symbol, whereas low-order circuitry 115 resolves the immediate least-significant bit (LSB). An immediate value of the MSB is used to select a set of ISI offsets used to resolve the LSB. Resolved values of the prior values of the MSB and LSB are then used to select the ISI offset for the immediate symbol.

Abstract: System and method for training and performing operations (e.g., read and write operations) on a double buffered memory topology. In some embodiments, eight DIMMs are coupled to a single channel. The training and operations schemes are configured with timing and signaling to allow training and operations with the double buffered memory topology. In some embodiments, the double buffered memory topology includes one or more buffers on a system board (e.g., motherboard).

Abstract: A memory component internally generates and stores the check bits of error detect and correct code (EDC). In a first mode, during a read transaction, the check bits are sent to the memory controller along with the data on the data mask (DM) signal lines. In a second mode, an unmasked write transaction is defined where the check bits are sent to the memory component on the data mask signal lines. In a third mode, a masked write transaction is defined where at least a portion of the check bits are sent from the memory controller on the data signal lines coincident with an asserted data mask signal line. By sending the check bits along with the data, the EDC code can be used to detect and correct errors that occur between the memory component and the memory controller.

Abstract: A DRAM includes at least four groups of memory cores and at least four memory access channel interfaces that, in a first mode, each respectively are to receive memory access commands, directed to a corresponding one of the groups of memory cores. One-half of the memory access channel interfaces are to, in a second mode, each respectively receive memory access commands, directed to a corresponding two of four of the groups of memory cores. The memory access channel interfaces to have electrical connection conductors that lie on opposing sides of at least one line of reflectional symmetry from a second one-half of the one-half of the at least four memory access channel interfaces.

Abstract: A first-in-first-out (FIFO) storage structure within an integrated-circuit component is loaded with qualification values corresponding to respective pairs of edges expected within a timing strobe signal transmitted to the integrated-circuit component. The qualification values are sequentially output from the FIFO storage structure during respective cycles of the timing strobe signal and a gate signal is either asserted or deasserted during the respective cycles of the timing strobe signal according to the qualification values output from the FIFO storage structure.

Abstract: Same sized blocks of data corresponding to a single read/write command are stored in the same memory array of a memory device, but using different formats. A first one of these formats spreads the data in the block across a larger number of memory subarrays (a.k.a., memory array tiles—MATs) than a second format. In this manner, the data blocks stored in the first format can be accessed with lower latency than the blocks stored in the second format because more data can be read from the array simultaneously. In addition, since the data stored in the second format is stored in fewer subarrays, it takes less energy to read a block stored in the second format. Thus, a system may elect, on a data block by data block basis, whether to conserve power or improve speed.

Abstract: Memory controllers, devices, modules, systems and associated methods are disclosed. In one embodiment, a memory controller is disclosed. The memory controller includes write queue logic that has first storage to temporarily store signal components of a write operation. The signal components include an address and write data. A transfer interface issues the signal components of the write operation to a bank of a storage class memory (SCM) device and generates a time value. The time value represents a minimum time interval after which a subsequent write operation can be issued to the bank. The write queue logic includes an issue queue to store the address and the time value for a duration corresponding to the time value.

Abstract: A memory module comprises a data interface including a plurality of data lines and a plurality of configurable switches coupled between the data interface and a data path to one or more memories. The effective width of the memory module can be configured by enabling or disabling different subsets of the configurable switches. The configurable switches may be controlled by manual switches, by a buffer on the memory module, by an external memory controller, or by the memories on the memory module.

Abstract: Multiple device stacks are interconnected in a ring topology. The inter-device stack communication may utilize a handshake protocol. This ring topology may include the host so that the host may initialize and load the device stacks with data and/or commands (e.g., software, algorithms, etc.). The inter-device stack interconnections may also be configured to include/remove the host and/or to implement varying numbers of separate ring topologies. By configuring the system with more than one ring topology, and assigning different problems to different rings, multiple, possibly unrelated, machine learning tasks may be performed in parallel by the device stack system.

Abstract: A method of transferring data between a memory controller and at least one memory module via a primary data bus having a primary data bus width is disclosed. The method includes accessing a first one of a memory device group via a corresponding data bus path in response to a threaded memory request from the memory controller. The accessing results in data groups collectively forming a first data thread transferred across a corresponding secondary data bus path. Transfer of the first data thread across the primary data bus width is carried out over a first time interval, while using less than the primary data transfer continuous throughput during that first time interval. During the first time interval, at least one data group from a second data thread is transferred on the primary data bus.

Abstract: Described are motherboards with memory-module sockets that accept legacy memory modules for backward compatibility or accept a greater number of configurable modules in support of increased memory capacity. The configurable modules can be backward compatible with legacy motherboards. Equipped with the configurable modules, the motherboards support memory systems with high signaling rates and capacities.

Abstract: A memory access command, column address and plurality of write data values are received within an integrated-circuit memory chip via external signaling links. In response to the memory access command, the integrated-circuit memory chip (i) decodes the column address to select address-specified sense amplifiers from among a plurality of sense amplifiers that constitute a sense amplifier bank, (ii) reads first data, constituted by a plurality of read data values, out of the address-specified sense amplifiers, and (iii) overwrites the first data within the address-specified sense amplifiers with second data constituted by one or more of the write data values and by one or more of the read data values.

Abstract: A system can include a processor having a secure mode and a non-secure mode, and a secure module configured to respond to tokens posted by the processor in the secure mode. Each token can identify a secure asset, and source and destination addresses within secure and public address spaces. The secure module can include a memory storing secure assets identifiable by the tokens and a memory access circuit to read data from source addresses and write processed data to destination addresses. The system can further include a cryptography engine configured to process the read data using identified secure assets. The secure module can respond to tokens posted in the non-secure mode. The memory can store, with each secure asset, a respective rule defining the address spaces where the memory access circuit may read and write data. The secure module can ignore tokens that do not satisfy respective rules.

Abstract: A gating signal for masking overhead transitions in a data-strobe signal is generated adaptively based on timing events in the incoming data-strobe signal itself to yield a gating window that opens and closes deterministically with respect to active edges of the data-strobe signal.

Abstract: A method of operating a memory controller is disclosed. The method includes transmitting data signals to a memory device over each one of at least two parallel data links. A timing signal is sent to the memory device on a first dedicated link. The timing signal has a fixed phase relationship with the data signals. A data strobe signal is driven to the memory device on a second dedicated link. Phase information is received from the memory device. The phase information being generated internal to the memory device and based on a comparison between the timing signal and a version of the data strobe signal internally distributed within the memory device. A phase of the data strobe signal is adjusted relative to the timing signal based on the received phase information.

Abstract: A signal transmitter circuit includes an output driver circuit configured to transmit a signal using a multi-level pulse amplitude modulation (PAM) scheme comprising a plurality of discreet signal levels. During operation, the output driver initiates a first transition of the signal to a first level of the multi-level PAM scheme from a second level of the multi-level PAM scheme, and initiates a second transition of the signal to the first level from a third level of the multi-level PAM scheme. The signal transmitter further includes a control circuit configured to control a slew rate of the signal transmitter circuit to cause the signal to reach a threshold voltage level at a first time, the first time occurring a first duration of time after the first transition is initiated, and to cause the signal to reach the threshold voltage level at a second time, the second time occurring the first duration of time after the second transition is initiated.