A system-on-a-chip (SoC) with one or more processors and other system components may have one or more peripheral component interconnect express (PCIe) physical connections between the processors and other system components to provide agent-to-agent communication. The agents on the communication fabric of the SoC may transmit data through the hardware PCIe interface where a transmitter device of an agent or digital logic component receives at least one data block for transmission and receives a flag corresponding to the at least one data block. The transmitter device may then send, via a PCIe physical layer, the received data blocks as a payload of a packet based on the flag, where the packet has a PCIe compliant header. The payload of the packet with the PCIe header may be entirely composed of these data blocks or flits from the agent.

A system includes a fabric switch including a motherboard, a baseboard management controller (BMC), a network switch configured to transport network signals, and a PCIe switch configured to transport PCIe signals; a midplane; and a plurality of device ports. Each of the plurality of device ports is configured to connect a storage device to the motherboard of the fabric switch over the midplane and carry the network signals and the PCIe signals over the midplane. The storage device is configurable in multiple modes based a protocol established over a fabric connection between the system and the storage device.

Methods, systems, and apparatuses for self-referencing sensing schemes are described. A cell having two transistors, or other switching components, and one capacitor, such as a ferroelectric capacitor, may be sensed using a reference value that is specific to the cell. The cell may be read and sampled via one access line, and the cell may be used to generate a reference voltage and sampled via another access line. For instance, a first access line of a cell may be connected to one read voltage while a second access line of the cell is isolated from a voltage source; then the second access line may be connected to another read voltage while the first access line is isolate from a voltage source. The resulting voltages on the respective access lines may be compared to each other and a logic value of the cell determined from the comparison.

A system comprises a processor coupled to a plurality of memory units. Each of the plurality of memory units includes a request processing unit and a plurality of memory banks. Each request processing unit includes a plurality of decomposition units and a crossbar switch, the crossbar switch communicatively connecting each of the plurality of decomposition units to each of the plurality of memory banks. The processor includes a plurality of processing elements and a communication network communicatively connecting the plurality of processing elements to the plurality of memory units. At least a first processing element of the plurality of processing elements includes a control logic unit and a matrix compute engine. The control logic unit is configured to access the plurality of memory units using a dynamically programmable distribution scheme.

Some embodiments relate to a link monitor for a switch having a PCIe-compliant interface. Some embodiments relate to an apparatus including a Peripheral Component Interconnect Express (PCIe)-compliant interface provided at a PCIe domain of a switch. The apparatus may also include a link monitor provided at a switching fabric of the switch that supports the PCIe domain of the switch. The link monitor to observe a factor-changing event of a state of a fabric link and obtain a value at least partially responsive to a weight computation, the weight computation for a factor associated with the factor-changing event. Related devices, systems and methods are also disclosed.

Techniques for transmitting data include identifying data to be transmitted, adding the data to a queue, and in response to a data session window being open: extracting the data from the queue; transmitting the extracted data to a transceiver via a transmitter; monitoring an amount of data in the queue and determining that the transmitter has transmitted the extracted data to the transceiver; and in response, instructing the transceiver to end the data session window early and transition to a lower power state.

A multi-plane, multi-protocol memory switch system is disclosed. In some embodiments, a memory switch includes a plurality of switch ports, the memory switch connectable to one or more root complex (RC) devices through one or more respective switch ports of the plurality of switch ports, and the memory switch connectable to a set of endpoints through a set of other switch ports of the plurality of switch ports, wherein the set includes zero or multiple endpoints; a cacheline exchange engine configured to provide a data-exchange path between two endpoints and to map an address space of one endpoint to an address space of another endpoint; and a bulk data transfer engine configured to facilitate data-exchange between two endpoints as a source-destination data stream, one endpoint being designated a source address and another endpoint being designated a destination address.

Techniques for transmitting data include identifying data to be transmitted; and in response to a data session window being open: transmitting the data to a transceiver via a transmitter; determining whether there is additional data to be transmitted and determining whether the transmitter has transmitted the data to the transceiver; and in response, instructing the transceiver to end the data session window early and transition to a lower power state.

The invention relates to an interconnection network for forward error correction encoders and decoders, including N input terminals, N output terminals, and M stages. Each stage includes switching elements having input pins and output pins. The input pins of the switching elements of the first stage are connected to the input terminals, and the output pins of the switching elements of the last stage are connected to the output terminals. The input and output pins of the switching elements of immediately successive stages are connected in a hardwired fashion so as to form a plurality of interconnection sub-networks for routing respective input values from respective output pins of the switching elements of the first stage to respective input pins of the switching elements of the last stage.

A system is disclosed. The system may include a processor, a storage device, and aa computational device. A plugin may be associated with a service on the computational device. A framework may be implemented in software and configured to be executed on the processor. The framework may include a receiver to receive a service request from an application configured to be executed on the processor and a service response from the plugin. The service request may identify a service requested by the application. The framework may also include a plugin selector to select the plugin based at least in part on the service. The framework may also include a transmitter to deliver the service request to the plugin and the service response to the application. The application may be agnostic to the plugin and the computational device.