Abstract: Embodiments disclosed herein include an electronic package. In an embodiment, the electronic package comprises a package substrate, and a die coupled to the package substrate. In an embodiment, the electronic package further comprises a stiffener on the package substrate surrounding the die. In an embodiment, the stiffener is a ring with one or more corner regions and one or more beams. In an embodiment, each beam is between a pair of corner regions, and the one or more corner regions have a first thickness and the one or more beams have a second thickness that is greater than the first thickness.

Abstract: The present disclosure relates to methods and apparatuses for execution of a neural network. An exemplary method can be implemented by a processing unit. The processing unit can include a command parser configured to dispatch commands and computing tasks and at least one core communicatively coupled with the command parser and configured to process the dispatched computing task. Each core can include a convolution unit, a pooling unit, at least one operation unit and a sequencer communicatively coupled with the convolution unit, the pooling unit, and the at least one operation unit and configured to distribute instructions of the dispatched computing task to the convolution unit, the pooling unit, and the at least one operation unit for execution.

Abstract: A radio frequency heating method includes: cladding an object to be heated by using a surrounding medium, and performing radio frequency heating operation on the object to be heated clad by the surrounding medium, in which the surrounding medium is an alcoholic solution. In the radio frequency heating method according to the embodiments of the present application, by using the alcoholic solution as the surrounding medium to clad the object to be heated, heating uniformity of the object to be heated in a radio frequency heating process may be improved, edge corner energy concentration effect in the radio frequency heating process may be fully avoided, and then quality of the object to be heated after heating may be ensured.

Abstract: The present disclosure relates to a PCR detection kit for rapidly identifying Salmonella of specific serotypes. The kit includes primers for detecting tcpS gene, the primers for detecting the tcpS gene including a forward primer having a nucleotide sequence as set forth in SEQ ID NO. 1 and a reverse primer having a nucleotide sequence as set forth in SEQ ID NO. 2. The kit can identify Salmonella of enteritidis, pullorum/gallinarum, and dublin serotypes rapidly and in a high throughput, which can be used as an auxiliary method for the conventional serotyping of Salmonella.

Abstract: The presented systems enable efficient and effective network communications. In one embodiment, a system comprises a parallel processing unit (PPU) included in a chip and a plurality of interconnects in an inter-chip network (ICN) configured to communicatively couple a plurality of PPUs that communicate with one another via the ICN. Corresponding communications are configured in accordance with routing tables. The routing tables can be stored and reside in registers of an ICN subsystem included in the PPU and include indications of minimum links available to forward a communication from the PPU and another PPU. Respective ones of the routing tables include indications of a correlation between the minimum links and respective ones of a plurality of egress ports that are available for communication coupling to the respective ones of PPUs that are possible destination PPUs. The routing tables can include indications of a single path between two PPUs per respective information communication flow.

Abstract: A system includes a high-bandwidth inter-chip network (ICN) that allows communication between neural network processing units (NPUs) in the system. For example, the ICN allows an NPU to communicate with other NPUs on the same compute node (server) and also with NPUs on other compute nodes (servers). Communication can be at the direct memory access (DMA) command level and at the finer-grained load/store instruction level. The ICN system and the programming model allows NPUs in the system to communicate without using a traditional network (e.g., Ethernet) that uses a relatively narrow and slow Peripheral Component Interconnect Express (PCIe) bus.

Abstract: Systems, methods, and devices for decentralized notification services are disclosed. In some aspects, the techniques described herein relate to: providing a distributed notification service configured to execute on each node of a plurality of nodes of a decentralized peer-to-peer platform; determining, by the distributed notification service of a first node, a notification definition, where the notification definition is from a catalog of notification definitions and included in a distributed application executing on the first node; sending, by the distributed notification service of a second node, a notification message, where the notification message is associated with the notification definition; receiving, by the distributed notification service of the first node, the notification message; and delivering, by the distributed notification service of the first node, the notification message in the distributed application executing on the first node.

Abstract: An electronic device includes remapping hardware, a processor, and a Northbridge IC. The remapping hardware converts a virtual address included in an unconverted DMA request into a physical address. The processor executes software to configure the remapping hardware. The Northbridge IC sends the physical address to the processor. When the software changes the configuration of the remapping hardware, the remapping hardware outputs a data draining request to the Northbridge IC. When the Northbridge IC receives the data draining request at a first time, the Northbridge IC suspends unconverted DMA requests after the first time until a second time, and outputs a first data draining response to the remapping hardware at the second time. The remapping hardware receives the first data draining response and notifies the processor that the data draining request has been completed.

Abstract: The present disclosure relates to a device for executing a convolutional neural network operation. The device comprises a first memory, a processing array comprising a plurality of processing strings, and a controller. The controller can be configured to fetch one or more batches of data into the first memory, regroup the one or more batches of data into multiple work items, wherein a first work item partially overlaps one or more work items among the multiple work items, and broadcast the multiple work items to the processing array, wherein the first work item is transferred to two or more processing strings of the processing array.

Abstract: The present disclosure provides a system comprising: a first group of computing nodes and a second group of computing nodes, wherein the first and second groups are neighboring devices and each of the first and second groups comprising: a set of computing nodes A-D, and a set of intra-group interconnects, wherein the set of intra-group interconnects communicatively couple computing node A with computing nodes B and C and computing node D with computing nodes B and C; and a set of inter-group interconnects, wherein the set of inter-group interconnects communicatively couple computing node A of the first group with computing node A of the second group, computing node B of the first group with computing node B of the second group, computing node C of the first group with computing node C of the second group, and computing node D of the first group with computing node D of the second group.

Abstract: An electronic device is provided. The electronic device includes a memory and an integrated circuit. The integrated circuit includes an address remapping unit. The memory includes multiple memory pages. The integrated circuit converts multiple virtual addresses into multiple physical addresses in sequence. The address remapping unit prefetches a first physical address corresponding to a first virtual address if a second virtual address exceeds a preset offset. The first virtual address is in a different memory page from the second virtual address. The second virtual address is currently processed. The multiple virtual addresses include the first and second virtual addresses.

Abstract: In one example, a controller circuit is configured to: receive a first measurement signal representing a power converter input voltage; receive a second measurement signal representing a power converter output voltage; receive an indication of whether a voltage across a switch is positive during a state change of the switch; based on the first and second measurement signals and the indication, determine: a charging interval; a first dead time interval after the charging interval; a discharging interval; and a second dead time interval after the discharging interval; and provide a first control signal at the first control output, and provide a second control signal at the second control output, in which the states of the first and second control signals vary across the charging interval, the first and second dead time intervals, and the discharging the interval.

Abstract: The present disclosure provides a method for syncing data of a computing task across a plurality of groups of computing nodes. Each group including a set of computing nodes A-D, a set of intra-group interconnects that communicatively couple computing node A with computing nodes B and C and computing node D with computing nodes B and C, and a set of inter-group interconnects that communicatively couple each of computing nodes A-D with corresponding computing nodes A-D in each of a plurality of neighboring groups. The method comprises syncing data at a computing node of the plurality of groups of computing nodes using inter-group interconnects and intra-group interconnects along four different directions relative to the node; and broadcasting synced data from the node to the plurality of groups of computing nodes using inter-group interconnects and intra-group interconnects along four different directions relative to the node.

Abstract: The present disclosure provides a method for syncing data of a computing task across a plurality of groups of computing nodes, each group comprising a set of computing nodes A-D, a set of intra-group interconnects that communicatively couple computing node A with computing nodes B and C and computing node D with computing nodes B and C, and a set of inter-group interconnects that communicatively couple a computing node A of a first group of the plurality of groups with a computing node A of a second group neighboring the first group, a computing node B of the first group with a computing node B of the second group, a computing node C of the first group with the computing node C of the second group, and a computing node D of the to first group with a computing node D of the second group, the method comprising: syncing across a first dimension of computing nodes using a first set of ring connections, wherein the first set of ring connections are formed using inter-group and intra-group interconnects that commun

Abstract: Two or more micropipettes are used to increase a microinjection throughput in an automated system for microinjecting adherent cells on a Petri dish. In the system, a motorized stage carrying the Petri dish sequentially visits the cells according to an optimized injection sequence. The sequence is selected by minimizing a total distance traveled by the motorized stage such that each cell is visited once by one of the micropipettes. Using multiple micropipettes advantageously reduces the minimized total distance over using a single micropipette to thereby increase the throughput. The optimized injection sequence is obtained by solving an equality-generalized traveling salesman problem. Each micropipette is mounted on a motorized micromanipulator. The motorized stage and motorized micromanipulators operate coordinately that each micromanipulator goes down or up during movement of the motorized stage to compensate for unevenness between a focus plane and a moving trajectory of the motorized stage.

Abstract: The present disclosure provides a system comprising: a first group of computing nodes and a second group of computing nodes, wherein the first and second groups are neighboring devices and each of the first and second groups comprising: a set of computing nodes A-D, and a set of intra-group interconnects, wherein the set of intra-group interconnects communicatively couple computing node A with computing nodes B and C and computing node D with computing nodes B and C; and a set of inter-group interconnects, wherein the set of inter-group interconnects communicatively couple computing node A of the first group with computing node A of the second group, computing node B of the first group with computing node B of the second group, computing node C of the first group with computing node C of the second group, and computing node D of the first group with computing node D of the second group.

Abstract: The present invention relates to the field of electrode material of a low temperature resistant supercapacitor, and in particular to a nickel foam-supported defective tricobalt tetroxide nanomaterial, a low temperature resistant supercapacitor and a preparation method thereof. The method includes the following steps: dissolving cobalt acetate in an ethylene glycol solution and stirring uniformly to obtain a pink transparent solution; adding hexadecyl trimethyl ammonium bromide to the pink transparent solution, and stirring until the hexadecyl trimethyl ammonium bromide dissolves to obtain a mixed solution; putting the mixed solution into a teflon-lined reactor, adding pretreated nickel foam for hydrothermal reaction, taking out the nickel foam after the reaction is completed, and ultrasonic cleaning the nickel foam repeatedly before drying; and heat-treating the nickel foam obtained after drying.

Abstract: The present disclosure provides a method for syncing data of a computing task across a plurality of groups of computing nodes, each group comprising a set of computing nodes A-D, a set of intra-group interconnects that communicatively couple computing node A with computing nodes B and C and computing node D with computing nodes B and C, and a set of inter-group interconnects that communicatively couple a computing node A of a first group of the plurality of groups with a computing node A of a second group neighboring the first group, a computing node B of the first group with a computing node B of the second group, a computing node C of the first group with the computing node C of the second group, and a computing node D of the first group with a computing node D of the second group, the method comprising: syncing across a first dimension of computing nodes using a first set of ring connections, wherein the first set of ring connections are formed using inter-group and intra-group interconnects that communica

Abstract: A conformable heat sink interface for an integrated circuit package comprises a mounting plate having a first surface and a deformable membrane having a portion bonded to a second surface of the plate. A cavity is between the second surface of the plate and the deformable membrane. A flowable heat transfer medium is within the cavity. The flowable heat transfer medium has a thermal conductivity of not less than 30 W/m K. The deformable membrane is to conform to a three-dimensional shape of an IC package and the mounting plate has a second surface that is to be adjacent to a heat sink base.