Abstract: A compute system includes an execution unit (e.g. of a CPU) with a memory controller providing access to a hybrid physical memory. The physical memory is “hybrid” in that it combines a cache of relatively fast, durable, and expensive memory (e.g. DRAM) with a larger amount of relatively slow, wear-sensitive, and inexpensive memory (e.g. flash). A hybrid controller component services memory commands from the memory controller component and additionally manages cache fetch and evict operations that keep the cache populated with instructions and data that have a high degree of locality of reference. The memory controller alerts the hybrid controller of available access slots to the cache so that the hybrid controller can use the available access slots for cache fetch and evict operations with minimal interference to the memory controller.

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 stacked semiconductor device is disclosed that includes a plurality of semiconductor dies. Each die has oppositely disposed first and second surfaces, with pads formed on each of the surfaces. A plurality of through-vias connect respective pads on the first surface to respective pads on the second surface. The through-vias include a first group of through-vias coupled to respective I/O circuitry on the semiconductor die and a second group of through-vias not coupled to I/O circuitry on the semiconductor die. The plurality of semiconductor dies are stacked such that the first group of through-vias in a first one of the plurality of semiconductor dies are aligned with respective ones of at least a portion of the second group of through-vias in a second one of the plurality of semiconductor dies.

Abstract: A system and/or an integrated circuit including: (a) a multiplier-accumulator execution pipeline including multiplier-accumulator circuits to process image data, using associated filter weights, via a plurality of multiply and accumulate operations and (b) first data format conversion circuitry including (i) inputs to receive filter weights of a plurality of sets of filter weights, wherein each set includes a plurality of filter weights each having a block-scaled fraction data format, (ii) conversion circuitry, coupled to the inputs, to convert the filter weights of each set from the block-scaled fraction data format to a floating point data format, and (iii) outputs to output the filter weights having the floating point data format. In operation, the multiplier-accumulator circuits of the pipeline are configured to perform the plurality of multiply and accumulate operations using (a) the image data and (b) the filter weights having the floating point data format.

Abstract: Memory devices and a memory controller that controls such memory devices. Multiple memory devices receive commands and addresses on a command/address (C/A) bus that is relayed point-to-point by each memory device. Data is received and sent from these devices to/from a memory controller in a point-to-point configuration by adjusting the width of each individual data bus coupled between the individual memory devices and the memory controller. Along with the C/A bus are clock signals that are regenerated by each memory device and relayed. The memory controller and memory devices may be packaged on a single substrate using package-on-package technology. Using package-on-package technology allows the relayed C/A signals to connect from memory device to memory device using wire bonding. Wirebond connections provide a short, high-performance signaling environment for the chip-to-chip relaying of the C/A signals and clocks from one memory device to the next in the daisy-chain.

Abstract: A memory controller and buffers on memory modules each operate in two modes, depending on the type of motherboard through which the controller and modules are connected. In a first mode, the controller transmits decoded chip-select signals independently to each module, and the motherboard data channel uses multi-drop connections to each module. In a second mode, the motherboard has point-to-point data channel and command address connections to each of the memory modules, and the controller transmits a fully encoded chip-select signal group to each module. The buffers operate modally to correctly select ranks or partial ranks of memory devices on one or more modules for each transaction, depending on the mode.

Abstract: An integrated circuit including a multiplier-accumulator execution pipeline including a plurality of multiplier-accumulator circuits to process the data, using filter weights, via a plurality of multiply and accumulate operations. The integrated circuit includes first conversion circuitry, coupled the pipeline, having inputs to receive a plurality of sets of data, wherein each set of data includes a plurality of data, Winograd conversion circuitry to convert each set of data to a corresponding Winograd set of data, floating point format conversion circuitry, coupled to the Winograd conversion circuitry, to convert the data of each Winograd set of data to a floating point data format. In operation, the multiplier-accumulator circuits are configured to perform the plurality of multiply and accumulate operations using the data of the plurality of Winograd sets of data from the first conversion circuitry and the filter weights, and generate output data based on the multiply and accumulate operations.

Abstract: A controller includes a link interface that is to couple to a first link to communicate bi-directional data and a second link to transmit unidirectional error-detection information. An encoder is to dynamically add first error-detection information to at least a portion of write data. A transmitter, coupled to the link interface, is to transmit the write data. A delay element is coupled to an output from the encoder. A receiver, coupled to the link interface, is to receive second error-detection information corresponding to at least the portion of the write data. Error-detection logic is coupled to an output from the delay element and an output from the receiver. The error-detection logic is to determine errors in at least the portion of the write data by comparing the first error-detection information and the second error-detection information, and, if an error is detected, is to assert an error condition.

Abstract: Bandwidth for information transfers between devices is dynamically changed to accommodate transitions between power modes employed in a system. The bandwidth is changed by selectively enabling and disabling individual control links and data links that carry the information. During a highest bandwidth mode for the system, all of the data and control links are enabled to provide maximum information throughout. During one or more lower bandwidth modes for the system, at least one data link and/or at least one control link is disabled to reduce the power consumption of the devices. At least one data link and at least one control link remain enabled during each low bandwidth mode. For these links, the same signaling rate is used for both bandwidth modes to reduce latency that would otherwise be caused by changing signaling rates. Also, calibration information is generated for disabled links so that these links may be quickly brought back into service.

Abstract: A memory component and a controller communicate commands and data with each other The commands to activate and then access data, and the data itself, are all communicated between a controller and the memory component at different times. The controller and memory component each calculate a respective error detecting code (EDC) values on the activate command information (e.g., bank address and row address) and store them indexed by the bank address. When the memory component is accessed, retrieved EDC values are combined with EDC values calculated from the access command information, and the data itself. The memory component transmits its combined EDC value to the controller for checking.

Abstract: In a chip-to-chip signaling system includes at least one signaling link coupled between first and second ICs, the first IC has an interface coupled to the signaling link and timed by a first interface timing signal. The second IC has an interface coupled to the signaling link and timed by a second interface timing signal that is mesochronous with respect to the first interface timing signal. The second IC further has phase adjustment circuitry that adjusts a phase of the second interface timing signal using a digital counter implemented with Josephson-junction circuit elements.

Abstract: Embodiments generally relate to a memory device. In one embodiment, the memory device includes a clock receiver circuit that receives an external clock signal and provides an internal clock signal. The memory device also includes a delay-locked loop circuit (DLL) having an input, and a circuit that receives the internal clock signal. The circuit selects which pulses of the internal clock signal are applied to the input of the DLL, such that no more than two clock pulses selected from at least three consecutive pulses of the external clock signal are applied to the input of the DLL during a predetermined interval.

Abstract: An integrated circuit includes a delay circuit and first and second interface circuits. The delay circuit delays a first timing signal by an internal delay to generate an internal timing signal. The first interface circuit communicates data to an external device in response to the internal timing signal. The second interface circuit transmits an external timing signal for capturing the data in the external device. An external delay is added to the external timing signal in the external device to generate a delayed external timing signal. The delay circuit sets the internal delay based on a comparison between the delayed external timing signal and a calibration signal transmitted by the first interface circuit.

Abstract: Memory controllers, devices, modules, systems and associated methods are disclosed. In one embodiment, a memory module includes a pin interface for coupling to a bus. The bus has a first width. The module includes at least one storage class memory (SCM) component and at least one DRAM component. The memory module operates in a first mode that utilizes all of the first width, and in a second mode that utilizes less than all of the first width.

Abstract: A memory cell includes a first resistive memory element, a second resistive memory element electrically coupled with the first resistive memory element at a common node, and a switching element comprising an input terminal electrically coupled with the common node, the switching element comprising a driver configured to float during one or more operations.

Abstract: An integrated circuit device is disclosed including core circuitry and interface circuitry. The core circuitry outputs in parallel a set of data bits, while the interface circuitry couples to the core circuitry. The interface circuitry receives in parallel a first number of data bits among the set of data bits from the core circuitry and outputs in parallel a second number of data bits. The ratio of the first number to the second number is a non-power-of-2 value.

Abstract: Memory controllers, devices, modules, systems and associated methods are disclosed. In one embodiment, an integrated circuit (IC) memory device is disclosed. The memory device includes an array of storage cells and command interface circuitry to receive an internal transfer command. In response to the internal transfer command, transfer logic reads data from a first portion of the array of storage cells, transfers the data as on-chip transfer data, and writes the on-chip transfer data to a second portion of the array of storage cells. In response to the command interface circuitry receiving an interrupt command, the transfer logic pauses the internal transfer operation, and carries out an unrelated memory access operation involving at least the first portion of the array of storage cells or the second portion of the array of storage cells.

Abstract: Disclosed is a memory system including a memory component having at least one tag row and at least one data row and multiple ways to hold a data group as a cache-line or cache-block. The memory system includes a memory controller that is connectable to the memory component to implement a cache and operable with the memory controller and the memory component in each of a plurality of operating modes including a first and second operating mode having differing addressing and timing requirements for accessing the data group. The first operating mode having placement of each of at least two ways of a data group in differing rows in the memory component, with tag access and data access not overlapped. The second operating mode having placement of all ways of a data group in a same row in the memory component, with tag access and data access overlapped.

Abstract: The embodiments described herein describe technologies for using the memory modules in different modes of operation, such as in a standard multi-drop mode or as in a dynamic point-to-point (DPP) mode (also referred to herein as an enhanced mode). The memory modules can also be inserted in the sockets of the memory system in different configurations.

Abstract: Described are memory modules that support different error detection and correction (EDC) schemes in both single- and multiple-module memory systems. The memory modules are width configurable and support the different EDC schemes for relatively wide and narrow module data widths. Data buffers on the modules support the half-width and full-width modes, and also support time-division-multiplexing to access additional memory components on each module in support of enhanced EDC.