67 results about "Main processing unit" patented technology

Machine learning model scalability with distributed multi-layer processing is disclosed herein. A method for processing and deploying machine learning models that enhances scalability and efficiency by executing a subset of a neural network on each of a plurality of interconnected processing units. The method involves partitioning compute tasks across these processing units to reduce latency, including broadcast and reduction processes for inputs and outputs. It also includes managing the allocation of samples in a batch to specific master processing units within the distributed arrangement and synchronizing computation between fully connected layers within each processing unit. Additionally, the method implements data reduction during the transfer of data across the processing units, wherein data is accumulated with a current processing unit's partial sum as it is transferred to the destination processing unit.

A control execution unit is connected both a main bus and a sub bus, and includes an execution unit and a selection unit. The execution unit performs vehicle control according to a selected manipulated variable being either a main manipulated variable from a main processing unit connected to the main bus or a sub manipulated variable from a sub processing unit connected to the sub bus. The selected manipulated variable is set to the main manipulated variable in an initial state, and the selection unit switches over the selected manipulated variable from the main manipulated variable to the sub manipulated variable when communication performed via the main bus satisfies a preset switchover condition.

The utility model discloses a measuring device. The measuring device comprises a base used for being installed on a measuring platform; the near end of the joint arm is connected to the base, and the far end of the joint arm is provided with a measuring device for the measuring device to move in the space so as to align to a point to be measured for measurement; the main control device is arranged on the base and comprises a power supply management unit which is provided with a battery and is powered by the battery when no external power supply exists, and a main processing unit which is electrically connected to the power supply management unit and is dormant when no external power supply exists; and the detection unit is electrically connected with the power supply management unit and the main processing unit so as to awaken the main processing unit to record an acceleration event when the acceleration event is detected. According to the measuring equipment disclosed by the invention, low-power-consumption work under the condition of no external power supply access such as transportation or carrying can be realized, acceleration events can be recorded, and the impact condition in the transportation or carrying process can be traced while energy consumption is saved.

A data scheduling method and computing equipment, the method is applied to a main processing unit, the main processing unit is in communication connection with a co-processing unit, the co-processing unit comprises a plurality of processing cores, and the method comprises: aligning target data of a target data volume according to a preset alignment boundary; determining a target level according to the aligned target data volume; determining a data segmentation parameter of the target data according to the target data volume and the target level; and the data segmentation parameters are sent to the co-processing unit, so that the co-processing unit schedules the processing core to execute calculation according to the data segmentation parameters. Therefore, the targeted data segmentation scheme is designed for the target data with different target data volumes to determine the data segmentation parameters, so that reasonable scheduling and distribution of a plurality of processing cores can be realized, and the rate of executing target binocular operator calculation by the system is improved.

The utility model relates to the technical field of network communication, and especially relates to a dual-channel fast switching network device comprising a main channel module which comprises a main network port, a main controller and a main processing unit; the standby channel module comprises a standby network port, a standby controller and a standby processing unit, and the standby channel module is in communication connection with the main channel module; the state monitoring module is arranged at the main channel module and is used for monitoring the main channel module in real time; and the switching control module is electrically connected with the state monitoring module and receives monitoring information of the state monitoring module on the main channel module, and the switching control module is electrically connected with the main channel module and the standby channel module and controls whether the main channel module and the standby channel module are switched or not. According to the utility model, when the abnormity of the main channel is detected, the switching operation is carried out, and the switching time is shortened to a sub-millisecond level, thereby meeting the requirements of application scenes with higher real-time requirements.

The present invention relates to a method and apparatus for writing data to a flash memory. The method for writing data to the flash memory is executed by a controller of an accelerator and includes: reading operation settings of a virtual vehicle from a static random access memory; when the operation settings indicate that first data of the virtual vehicle needs to go through an intermediate processing stage, setting a redundant array of independent disks (RAID) engine according to an intermediate parameter set stored in the static random access memory; when the operation settings indicate that the first data of the virtual vehicle does not need to go through the intermediate processing stage, or the first data of the virtual vehicle has already gone through the intermediate processing stage, determining whether second data of the virtual vehicle needs to go through a post-processing stage according to the operation settings; and when the operation settings indicate that the second data of the virtual vehicle needs to go through the post-processing stage, sending a programming index to a data access engine according to a post-processing parameter set stored in the static random access memory. Through the setting of a dedicated accelerator, the main processing unit does not need to supervise the operation of the entire data write data stream, improving the overall performance of the system.

A display apparatus includes: a video processing unit processing a video signal; a graphics processing unit processing a graphics signal; a mixing unit mixing video corresponding to the processed video signal and graphics corresponding to the processed graphics signal; a display unit outputting the mixed video and graphics; and a main processing unit configured to control the video processing unit, the graphics processing unit, the mixing unit and the display unit.

The invention relates to the technical field of the Internet of Things and the Internet, and discloses a low-power-consumption control method and device for an embedded system. The method is applied to equipment comprising a main processing unit, an auxiliary processing unit, a low-power-consumption monitoring module and a nonvolatile storage unit. The main processing unit and the auxiliary processing unit are connected through a communication interface and share the storage unit, and the auxiliary processing unit is connected with the low-power-consumption monitoring module. The main processing unit generates operation parameters with verification information, and the auxiliary processing unit verifies and stores the operation parameters and initializes the monitoring module; then the main unit sends running state data, and the auxiliary unit judges a sleep condition and generates an instruction in combination with the wake-up threshold value; the main unit responds to the instruction and stores and feeds back key data in a layered mode, and the auxiliary unit cuts off power supply after verification and enters a low-power-consumption standby mode. According to the scheme, a traditional sleep mode is broken through, standby power consumption is remarkably reduced, and the long-endurance requirement of the Internet of Things equipment is met.

The present invention describes a multi-tiered parallel array computer consisting of a main processing unit and its software manager connected to an array of nodes consisting of node processing devices with associated software managers wherein each node has an attached secondary array consisting of a combination of processing devices and or storage devices. Means are provided using the software managers that allow varying modes of data access include exclusive data access (or “Disjoint Access”) to the secondary array storage devices by the main processing unit or any or node processing devices. Also, enhancing speed devices such as cross-point switches may be used to provide maximize processor efficiency. The present invention is particularly suited for artificial intelligence machine learning computation or complex edge computing where multiple processors are used to provide low latency processing of large data sets and where the processors and data storage are housed in a single server case.

A method and a system records machining information associated with insulating panels manufactured in a continuous machining line. Machining units carry out machining operations on a panel body moved along the line, between an inlet and an outlet, from a start-production unit at the line inlet. The insulating panels are manufactured by making predetermined cuts in the panel body, at the line outlet. Each machining unit is controlled by a respective electronic control unit. The method generates and prints unique graphic representation identification codes, or Data Matrix codes. A panel body portion to be machined is identified. The Data Matrix codes printed on the panel body to be machined are identified when the panel body portions pass through the machining unit. Machining information is acquired and associated with the detected Data Matrix code. The acquired machining information and associated code are sent to a main processing unit connected.

The present invention relates to a computer-readable storage medium, a method and an apparatus for writing data into a flash memory. The method for writing data into the flash memory is executed by a main processing unit of a flash memory controller and includes: generating a front-end parameter set, a middle-end parameter set, and a back-end parameter set for each item of data writing; transmitting the front-end parameter set of each item to a routing engine so that the routing engine drives a host interface according to the content of the front-end parameter set; transmitting the middle-end parameter set of each item to an accelerator so that the accelerator drives a redundant array of independent disks (RAID) engine according to the content of the middle-end parameter set; and transmitting the back-end parameter set of each item to the accelerator so that the accelerator drives a data access engine according to the content of the back-end parameter set. Through the setting of an exclusive routing engine and accelerator, the main processing unit does not need to supervise the operation of the entire data writing data stream, improving the overall performance of the system.

The invention concerns a method of detection of a cyberattack event in a ground transportation vehicle, the method comprising at least one iteration of the following steps: - receiving (10), from a sensor, at least one datum representing the opening of a hatch of a maintenance compartment of said ground transportation vehicle, said maintenance compartment comprising at least one digital equipment connected to a main processing unit of said ground transportation vehicle by using a communication network; - obtaining (20) at least one datum representing a state of said at least one digital equipment; - determining (30) that said at least one datum representing the state of said at least one digital equipment and said at least one datum representing the opening of said hatch are in discordance with an operation state of said ground transportation vehicle; and - raising (40) at least one alarm.

Machine learning model scalability with distributed multi-layer processing is disclosed herein. A method for processing and deploying machine learning models that enhances scalability and efficiency by executing a subset of a neural network on each of a plurality of interconnected processing units. The method involves partitioning compute tasks across these processing units to reduce latency, including broadcast and reduction processes for inputs and outputs. It also includes managing the allocation of samples in a batch to specific master processing units within the distributed arrangement and synchronizing computation between fully connected layers within each processing unit. Additionally, the method implements data reduction during the transfer of data across the processing units, wherein data is accumulated with a current processing unit's partial sum as it is transferred to the destination processing unit.

The present invention provides a dual-system hot standby redundant BTM device, which is connected to an ATP main control unit, and includes a hot standby A-system BTM and a B-system BTM; wherein, for any system BTM: it includes an antenna unit and a main processing unit, the main processing unit is respectively connected to the antenna unit and the ATP main control unit, and is respectively connected to the two-system antenna units through a power supply enabling cross control line of the system, and can receive a control command of the ATP main control unit, and control any system antenna unit to power on or power off through the enabling cross control line according to the control command; the powered-on antenna unit communicates with the ground transponder passing by, receives the transponder message sent by the transponder and sends it to the main processing unit after processing, and the main processing unit sends it to the ATP main control unit after processing. The present invention can control the antenna units to be used alternately, prolong the service life of high-power devices, and improve the reliability and availability of BTM equipment.

The present invention discloses a communication control server based on packet domain communication, and its system and method. The communication control server comprises: a core logic processing communication unit, which includes a main processing unit for core logic processing and a communication unit for information exchange with the outside world; a network interaction unit connected to the communication unit, which is connected to the core logic processing communication unit. The network interaction unit is respectively connected to the dispatching centralized control system, the train control center, and the on-board equipment, and exchanges information with the on-board equipment based on a packet domain wireless network. The communication unit transmits a main check word to a safety supervision module. When the safety supervision module determines that the main check word is incorrect, a safety relay driver disconnects the external data transmission channel between the communication unit and the network interaction unit. Advantageously, the communication control server exchanges information with the outside world based on a packet domain wireless network, eliminating the need for an additional ISDN wireless communication server, thereby simplifying the overall structure and cost.

The invention relates to the technical field of substation secondary equipment, and discloses a hardware module interchange method, system, equipment and medium of a universal protection measurement and control device, and the method comprises the following steps: when a new hardware module is inserted into a physical slot of the protection measurement and control device, the new hardware module actively reports metadata to a main processing unit of the device; the configuration management service performs drive matching search in a universal drive library, and dynamically loads a corresponding drive program for the new hardware module; allocating system resources required by operation, and performing initial configuration on the new hardware module through the driving program; and reading the original data generated by the new hardware module, converting the original data into standardized engineering value data, and issuing the standardized engineering value data to a real-time data bus. According to the invention, the complexity and cost of spare part management and on-site operation and maintenance are greatly reduced, the research and development and verification period brought by element iteration is shortened, and the stability and reliability of system operation are improved.

A display driving apparatus and a display apparatus are provided. The display apparatus includes: a video processing unit that processes a video signal; a graphic processing unit that processes the graphic signal; a mixing unit that mixes a video corresponding to the processed video signal and a graph corresponding to the processed graph signal; the display unit is used for outputting the mixed video and graph; and a main processing unit configured to control the video processing unit, the graphics processing unit, the mixing unit, and the display unit.

In a device control system, a user device, a sequence generation device, and an information processing device (400) are connected via a network. An information processing device (400) is provided with a main processing unit (422). The main processing unit (422) acquires an evaluation formula in which one or more functions indicating a process executed by a device or a person and having a start condition, an execution parameter, and an end condition as input variables are arranged. The main processing unit (422) displays information relating to the process for each process on the basis of one or a plurality of functions included in the acquired evaluation formula.

Embodiments of the present application provide a system starting method and device, a storage medium and an electronic device, wherein the system starting method comprises: in the case of detecting that a first system starts, determining whether the first system is a cold start; in the case of determining that the first system is a cold start, detecting a central processing unit by a memory self-checking algorithm to determine whether there is a bad core in the central processing unit, wherein the bad core is used to indicate a damaged slave processing unit in the central processing unit, the first system starts at the same time as a master processing unit in the central processing unit, and the first system is used to control the master processing unit; in the case of there being a bad core in the central processing unit, isolating the bad core in the central processing unit, and starting a second system and other slave processing units in the central processing unit, wherein the second system is used to control the master processing unit and the other slave processing units, and the other slave processing units are slave processing units in the central processing unit except the bad core.

Digital number plate systems that include electronic displays and can integrate with broader vehicle systems, as well as associated methods, apparatuses, and systems, are disclosed herein. The digital number plate system can be comprised of a digital number plate that includes an electronic display that can be controlled by a mobile computing device. To do this the digital number plate includes a main processing unit with a communication module. The main processing unit processes and controls display data received by the electronic display. The communication module facilitates communication with the mobile computing device and with a host server and / or network.

The invention discloses an aircraft height measuring method, device and equipment and a storage medium, is applied to a wireless altimeter, and relates to a main processing unit, a monitoring processing unit, a radio frequency processing unit and an interface unit. The radio frequency processing unit sends out a transmitting instruction, the interface unit sends out a transmitting signal according to the transmitting instruction and receives an echo signal, and the echo signal is transmitted back to the radio frequency processing unit to be processed and then sent to the main processing unit and the monitoring processing unit. The main processing unit and the monitoring processing unit respectively process the echo signal to obtain first height data and second height data, compare the first height data with the second height data, judge whether a height error threshold value is met or not, and send a height output instruction when the height error threshold value is met; the main processing unit outputs first height data as the current height through the interface unit according to the height output instruction. The main processing unit and the monitoring processing unit respectively process echo signals and compare data, so that errors can be effectively reduced, and the measurement accuracy is improved.

The present invention discloses an instruction execution method of a digital signal processor and a digital signal processor, the instruction execution method includes: obtaining a target instruction and a peripheral clock enable signal of a corresponding peripheral unit; determining the instruction type of the target instruction through the instruction structure of the target instruction and a preset instruction setting condition; determining the signal state of the peripheral clock enable signal; executing a priority processing strategy when the instruction type is a priority instruction and the signal state is an enable state, the priority processing strategy includes: sending a companion core start signal to a priority processing unit, and sending a main core start signal to a main processing unit to start the priority processing unit to load and process the target instruction, and the main processing unit simultaneously processes the remaining instructions in the instruction set through a multi-stage pipeline sequence. The instruction execution method of the embodiment of the present invention can improve the instruction priority, accelerate the operation and processing of proprietary instructions, thereby improving the response speed and the processing capacity of the digital signal processor.

Machine learning model scalability with distributed multi-layer processing is disclosed herein. A method for processing and deploying a machine learning model is disclosed that improves scalability and efficiency by executing a subset of neural networks on each of a plurality of interconnected processing units. The method involves dividing computational tasks across these processing units to reduce latency, the computational tasks including broadcast and simplified processes for inputs and outputs. The method also includes managing allocation of samples in a batch to a particular primary processing unit within the distributed arrangement, and synchronizing computations between fully connected layers within each processing unit. Further, the method implements data simplification during transfer of data across the processing units, where the data is accumulated with a partial sum of the current processing unit as it is transferred to the destination processing unit.

The present application relates to an ultrafast laser parallel processing system, and to the field of laser precision processing technology. The ultrafast laser parallel processing system includes a main processing unit, and the main processing unit includes: a first lifting and conveying module, a double-layer conveying module, a second lifting and conveying module, a connecting and conveying module, a two-way conveying line module, an identification module, a frame, a laser processing module, and an image detection module connected in sequence along a first direction. In the process of processing the workpiece, the above-mentioned ultrafast laser parallel processing system improves the efficiency of the processing line, saves time, and at the same time improves the efficiency during processing, thereby realizing automated low-cost, large-scale production.

A primary processing unit includes queues configured to store commands prior to execution in corresponding pipelines. The primary processing unit also includes a first table configured to store entries indicating dependencies between commands that are to be executed on different ones of a plurality of processing units that include the primary processing unit and one or more secondary processing units. The primary processing unit also includes a scheduler configured to release commands in response to resolution of the dependencies. In some cases, a first one of the secondary processing units schedules the first command for execution in response to resolution of a dependency on a second command executing in a second one of the secondary processing units. The second one of the secondary processing units notifies the primary processing unit in response to completing execution of the second command.

The present invention provides a communication system, a processing unit, a server and an uplink data processing unit. The communication system includes a server and a first processing unit; the first processing unit includes a plurality of forward transmission ports, a plurality of return transmission ports, a plurality of first server communication interfaces and a main processing unit; the main processing unit is used to perform low-level protocol layer processing on the received forward transmission data, and is also used to perform preprocessing on the received return transmission data; the server is used to perform high-level protocol layer processing on the received return transmission data and / or forward transmission data; the high-level protocol layer processing is the data processing of the protocol layer above the physical layer and the physical layer in the protocol stack; the low-level protocol layer processing is the data processing of the bottom layer of the physical layer in the protocol stack. The combination of the server and the first processing unit not only realizes the data processing of the complete communication protocol stack, but also has the ability to interact with other units for data. The communication system has a simple architecture, and the difficulty and cost of research and development are relatively low.

Embodiments of this application relate to the field of Bluetooth technologies, and provide a Bluetooth data processing method, a terminal device, and a readable storage medium. The terminal device includes a main processing unit, an auxiliary processing unit, and a Bluetooth chip. The Bluetooth data processing method is applied to a terminal device, and includes: The Bluetooth chip receives first Bluetooth data, and sends the first Bluetooth data to the auxiliary processing unit if the Bluetooth chip determines that the first Bluetooth data meets a preset condition; and the auxiliary processing unit processes the first Bluetooth data. In embodiments of this application, some Bluetooth services are deployed on the auxiliary processing unit for processing. The Bluetooth chip has a Bluetooth data filtering capability, and sends Bluetooth data that meets the preset condition to the auxiliary processing unit for processing. This shortens working time of the main processing unit, and reduces power consumption of the terminal device.

The invention provides a hot backup error prevention system of a UALink protocol switch and a hot backup error prevention method thereof. The hot backup error prevention system of the UALink protocol switch comprises a main processing unit, a hot backup error prevention unit and a hot backup error prevention unit, the hot backup processing unit and the main processing unit synchronously receive the configuration data from the cluster controller and perform synchronous operation; the comparator is used for comparing whether the operation results of the main processing unit and the hot backup processing unit are consistent or not; the monitoring module is used for monitoring whether a fatal error exists or not; the discriminator is used for judging whether the state of the hot backup processing unit is normal or not; and the switching judgment module is used for judging whether to output a switching signal according to the output results of the comparator, the monitoring module and the discriminator so as to switch to the hot backup processing unit for error processing.

A control system for an autonomous vehicle includes a time sensitive network switch and a control system interface. The control system interface includes a primary processing unit, a secondary processing unit, and a fault detection module. The control system receives a trajectory of an autonomous vehicle. The fault detection module detects a fault condition with the primary processing unit and prevents the primary processing unit from transmitting a primary control signal in response to the fault condition. The primary control signal is for autonomous control of the autonomous vehicle.

The embodiment of the invention provides a data processing method, a main processing unit, a chip system, a device, equipment, a storage medium and a program product. The method includes, at a main processing unit, in response to receiving first voice data for a digital assistant, transmitting the first voice data to a first co-processing unit associated with the digital assistant to perform processing on the first voice data, the main processing unit being communicatively connected to a plurality of co-processing units including the first co-processing unit, the plurality of co-processing units are respectively configured to execute different processing operations; receiving the processed first voice data from the first co-processing unit; determining encoded speech data based on the processed first speech data; and sending the encoded voice data to a digital assistant. Therefore, the relevant processing of the digital assistant can be executed through the proper co-processing unit, so that the computing power of each processing unit is fully utilized, and the power consumption of the whole machine is reduced.