Abstract: A method for allocating resources for a machine learning model is disclosed. A machine learning model to be executed on a special purpose machine learning model processor is received. A computational data graph is generated from the machine learning model. The computational dataflow graph represents the machine learning model which includes nodes, connector directed edges, and parameter directed edges. The operations of the computational dataflow graph is scheduled and then compiled using a deterministic instruction set architecture that specifies functionality of a special purpose machine learning model processor. An amount of resources required to execute the computational dataflow graph is determined. Resources are allocated based on the determined amounts of resources required to execute the machine learning model represented by the computational dataflow graph.

Abstract: A system contains a network of processors arranged in a plurality of nodes. Each node comprises a respective plurality of processors connected via local links, and different nodes are connected via global links. The processors of the network communicate with each other to establish a global counter for the network, enabling deterministic communication between the processors of the network. A compiler is configured to explicitly schedule communication traffic across the global and local links of the network of processors based upon the deterministic links between the processors, which enable software-scheduled networking with explicit send or receive instructions executed by functional units of the processors at specific times, to establish a specific ordering of operations performed by the network of processors. In some embodiments, the processors of the network of processors are tensor streaming processors (TSPs).

Abstract: A power conversion apparatus comprises a voltage bus has a first rail, a second rail, and a plurality of switch legs coupled between the first rail and the second rail, each switch leg has a switch assembly. A first resonant assembly is coupled in series with a first transformer assembly, and a second resonant assembly coupled in series with a second transformer assembly. The first resonant assembly and the first transformer assembly are coupled between a first switch leg of the plurality of switch legs and a second switch leg of the plurality of switch legs. The second resonant assembly and the second resonant assembly are coupled between the second switch leg and a third switch leg of the plurality of switch legs.

Abstract: A circuit comprises an input register configured to receive an input vector of elements, a control register configured to receive a control vector of elements, wherein each element of the control vector corresponds to a respective element of the input vector, and wherein each element specifies a permutation of a corresponding element of the input vector, and a permute execution circuit configured to generate an output vector of elements corresponding to a permutation of the input vector. Generating each element of the output vector comprises accessing, at the input register, a particular element of the input vector, accessing, at the control register, a particular element of the control vector corresponding to the particular element of the input vector, and outputting the particular element of the input vector as an element at a particular position of the output vector that is selected based on the particular element of the control vector.

Abstract: A shingle coating asphalt composition is provided that is produced from a paving grade asphalt. The asphalt composition comprises a paving-grade asphalt that has been modified with one or more polymer additives; and a secondary additive comprising one or more of a viscosity reducing agent, a wax, a salt of a fatty acid ester, and an amide of a fatty acid. The shingle coating asphalt coating composition is used to make a shingle. The shingle includes a substrate, the asphalt, and roofing granules.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for computing a layer output for a convolutional neural network layer, the method comprising: receiving a plurality of activation inputs; forming a plurality of vector inputs from the plurality of activation inputs, each vector input comprising values from a distinct region within the multi-dimensional matrix; sending the plurality of vector inputs to one or more cells along a first dimension of the systolic array; generating a plurality of rotated kernel structures from each of the plurality of kernel; sending each kernel structure and each rotated kernel structure to one or more cells along a second dimension of the systolic array; causing the systolic array to generate an accumulated output based on the plurality of value inputs and the plurality of kernels; and generating the layer output from the accumulated output.

Abstract: An assembly, has: a rotor of a rotary internal combustion engine, the rotor having two end faces and a peripheral face between the two end faces, the peripheral face defining apex portions circumferentially-spaced from one another, the rotor having seals, the seals including: apex seals extending between the two end faces and each located at a respective one of the apex portions, end seals engaging ends of the apex seals, and arc seals located on the two end faces and extending between the apex portions; and a tool engaged to the rotor and abutting the seals, the tool having: two cover plates, the rotor located between the two cover plates, the two cover plates abutting the end seals and the arc seals; and posts secured to one or more of the two cover plates and extending between the two cover plates, the posts abutting the apex seals.

Abstract: A system contains a network of processors arranged in a plurality of nodes. Each node comprises a respective plurality of processors connected via local links, and different nodes are connected via global links. The processors of the network communicate with each other to establish a global counter for the network, enabling deterministic communication between the processors of the network. A compiler is configured to explicitly schedule communication traffic across the global and local links of the network of processors based upon the deterministic links between the processors, which enable software-scheduled networking with explicit send or receive instructions executed by functional units of the processors at specific times, to establish a specific ordering of operations performed by the network of processors. In some embodiments, the processors of the network of processors are tensor streaming processors (TSPs).

Abstract: A shingle coating asphalt composition is provided that is produced from a paving grade asphalt. The asphalt composition comprises a paving-grade asphalt that has been modified with one or more polymer additives; and a secondary additive comprising one or more of a viscosity reducing agent, a wax, a salt of a fatty acid ester, and an amide of a fatty acid. The shingle coating asphalt coating composition is used to make a shingle. The shingle includes a substrate, the asphalt, and roofing granules.

Abstract: Various embodiments include heat and volatile-organic-compounds detecting systems. In one example, the heat-detecting system includes at least one heat sensor mounted externally to a device, such as a local power-box (LPB). The heat sensor has an area-of-detection to detect heat emitted from at least one face of the LPB at one or more locations. The heat-detecting system also includes a high-absorptance infrared-collector (HAIC) formed within the LPB to collect excessive heat generated by a component within the LPB. The excessive heat is correlated to a pre-determined temperature level, and a temperature of the collected excessive heat is measured by the heat sensor. Each of the heat sensor and the HAIC are coupled to a control module. Other apparatuses, designs, and methods are disclosed.

Abstract: A circuit for transposing a matrix comprising reversal circuitry configured, for each of one or more diagonals of the matrix, to receive elements of the matrix in a first vector and generate a second vector that includes the elements of the matrix in an order that is a reverse of an order of the elements of the matrix in the first vector, and rotation circuitry configured, for each of the one or more diagonals of the matrix, to determine a number of positions by which to rotate the elements of the matrix in the second vector, receive the second vector of elements of the matrix, and generate a third vector that includes the elements of the matrix in the second vector in an order that is a rotation of the elements of the matrix in the second vector by the determined number of positions.

Abstract: Various embodiments include heat and volatile-organic-compounds detecting systems. In one example, the heat-detecting system includes at least one heat sensor mounted externally to a device, such as a local power-box (LPB). The heat sensor has an area-of-detection to detect heat emitted from at least one face of the LPB at one or more locations. The heat-detecting system also includes a high-absorptance infrared-collector (HAIC) formed within the LPB to collect excessive heat generated by a component within the LPB. The excessive heat is correlated to a pre-determined temperature level, and a temperature of the collected excessive heat is measured by the heat sensor. Each of the heat sensor and the HAIC are coupled to a control module. Other apparatuses, designs, and methods are disclosed.

Abstract: A polymer masonry unit is disclosed that can include a polymer added to a quarry byproduct to manufacture a quality brick unit. The present disclosure solves the technological problem of providing a structurally sound brick or concrete alternative without the need for kiln firing, using traditionally unusable waste material. By combining quarry byproduct and a polymer, a polymer masonry unit can be fabricated having compressive strength and architectural utility. In one exemplary embodiment, fiber elements can be added to the byproduct and polymer mixture to increase structural stability. The present disclosure improves the performance of the system itself by providing a basic block or brick unit using an environmentally responsible manufacturing process that reduces cost and waste. The manufacturing process includes a polymer/base material that can be poured into molds that cures over a predetermined period, without the need for kiln firing.

Abstract: A control gate for a channel or pond adapted to be installed across a channel for liquids. The control gate includes: at least two panels hingedly connected together, the panels having opposing sides in sealing engagement with side walls of the channel, a lowermost panel being in sealing engagement with the floor of the channel. The panels are substantially vertically aligned in a fully closed position of the control gate. At least one first lifting element is connected to the top of an uppermost panel and is adapted to lower and raise the uppermost panel relative to its hinged connection in a concertina manner to provide an overshot position for the control gate. At least one second lifting element is connected to the bottom of the lowermost panel to raise and lower the lowermost panel from the floor of the channel to provide an undershot position for the control gate.

Abstract: A portable display apparatus for beverage aeration includes a floating compartment, an aerating system, a battery, a chipset, a visual display system, and a power button. The floating compartment includes a first half and a second half as the first half and the second half are hermetically connected to each other. The aerating system includes a pump, at least one inlet port, and a plurality of outlet ports. The pump is mounted within the floating compartment. The inlet port is integrated into the first half. The plurality of outlet ports is integrated into the first half and the second half. The inlet port is in fluid communication with the plurality of outlet ports through the pump. The battery and the chipset are mounted within the floating compartment. The visual display system is mounted onto the floating compartment.

Abstract: A circuit comprises an input register configured to receive an input vector of elements, a control register configured to receive a control vector of elements, wherein each element of the control vector corresponds to a respective element of the input vector, and wherein each element specifies a permutation of a corresponding element of the input vector, and a permute execution circuit configured to generate an output vector of elements corresponding to a permutation of the input vector. Generating each element of the output vector comprises accessing, at the input register, a particular element of the input vector, accessing, at the control register, a particular element of the control vector corresponding to the particular element of the input vector, and outputting the particular element of the input vector as an element at a particular position of the output vector that is selected based on the particular element of the control vector.

Abstract: A printing system comprises one or more printheads arranged to eject ink to a deposition region. Print media are held by vacuum suction against a movable support surface, which transports the print media though the deposition region. The vacuum suction is communicated from a vacuum source to the movable support surface via a vacuum plenum. An airflow control system comprises one or more dampers arranged in the vacuum plenum to control airflow between the vacuum source and the movable support surface. The dampers have an adjustable impedance to airflow through the damper. The airflow control system is configured to adjust the impedance of the damper based on a detected condition of the printing system. A detected condition of the printing system may include, for example, a sensed pressure in the vacuum plenum.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for computing a layer output for a convolutional neural network layer, the method comprising: receiving a plurality of activation inputs; forming a plurality of vector inputs from the plurality of activation inputs, each vector input comprising values from a distinct region within the multi-dimensional matrix; sending the plurality of vector inputs to one or more cells along a first dimension of the systolic array; generating a plurality of rotated kernel structures from each of the plurality of kernel; sending each kernel structure and each rotated kernel structure to one or more cells along a second dimension of the systolic array; causing the systolic array to generate an accumulated output based on the plurality of value inputs and the plurality of kernels; and generating the layer output from the accumulated output.

Abstract: An apparatus is disclosed that includes a switchable optical filter having a layer of switchable material; a light source providing light of a wavelength that causes the switchable material to transition from a faded state to a dark state, or a dark state to a faded state; and a switch for controlling activation of the light source.

Abstract: A shingle coating asphalt composition is provided that is produced from a paving grade asphalt. The asphalt composition comprises a paving-grade asphalt that has been modified with one or more polymer additives; and a secondary additive comprising one or more of a viscosity reducing agent, a wax, a salt of a fatty acid ester, and an amide of a fatty acid. The shingle coating asphalt coating composition is used to make a shingle. The shingle includes a substrate, the asphalt, and roofing granules.