38 results about "Memory bus" patented technology

The invention discloses a high-speed snapshot system and method, and relates to the technical field of image acquisition. The system comprises a trigger detection module, an image acquisition module, a processing and control module, a DDR cache chip and a storage and transmission module. The image acquisition module adopts a high-speed CMOS (Complementary Metal-Oxide-Semiconductor Transistor) image sensor and works in a high-frame-rate mode which is not lower than 120 frames per second under a preset resolution ratio; the processing and control module adopts a system-level chip integrated with an image signal processor and a hardware encoder; the DDR cache chip is connected with the processing and control module through a memory bus and serves as a high-speed data buffer pool. Through cooperative work of a high-frame-rate sensor and a hardware-level encoder and cooperation of independent DDR cache, clear, continuous and frame-loss-free snapshot of high-speed moving targets such as flying insects is realized, and the problems of image blurring and incomplete data caused by insufficient frame rate and limited data processing capability in the prior art are solved. The system is high in integration level, flexible in deployment and suitable for field monitoring and laboratory research.

The invention discloses a universal control module, which belongs to the technical field of industrial field control and comprises a Loongson 2K0300 processor, a DDR4 (Double Data Rate 4) memory bus, a CAN (Controller Area Network) bus interface, a network PHY (Physical Layer) chip, an Ethernet interface, an Ethercat slave station interface, an Ethercat controller, a network transformer, an RS232 transceiver, a transceiver LTM2881 and an LCD (Liquid Crystal Display) interface. The Loongson 2K0300 processor comprises a memory controller, a CAN controller, an Ethernet controller and a UART controller. And the network PHY chip is connected with the Ethernet interface. And the Loongson 2K0300 processor, the Ethercat controller, the network transformer and the Ethercat slave station interface are sequentially communicated with one another. And the UART controller is communicated with the RS232 transceiver. And the UART controller is connected with the transceiver LTM2881 through a pin 1 and a pin 2 of the UART. The universal control module provided by the invention not only can be used as a subsystem single machine to realize corresponding functions in a relatively complex comprehensive control system, but also can independently realize the goal of comprehensive control of the system in small and medium-sized control systems, and realizes the balance of cost and power consumption and the high-efficiency cost ratio of simplicity and easiness in use.

An integrated circuit includes a memory bus coupled to a random access memory (RAM). A plurality of ports enable hardware logic to access the RAM over the memory bus. A central processing unit (CPU) is coupled to the memory bus. A first hardware logic module is coupled to a first port, of the plurality of ports, to share access to the memory bus with the CPU. A plurality of hardware registers, coupled to the plurality of ports, include a first hardware register to store a number of accesses, by the first hardware logic module, to the memory bus over the first port.

The application relates to a B+ tree index construction method based on ZNS SSD cold and hot data dynamic placement, which comprises a persistent memory PM, a dynamic random access memory DRAM and a ZNS SSD hard disk; the persistent memory PM and the dynamic random access memory DRAM are directly connected with a CPU memory bus and have the ability of being accessed by the CPU by bytes; internal nodes of the B+ tree are stored in the dynamic random access memory DRAM, wherein the internal nodes only store index information for positioning leaf nodes; pseudo leaf nodes of the B+ tree are stored in the persistent memory PM, wherein the pseudo leaf nodes only store key index information of the leaf nodes; leaf nodes for actually storing key value pairs in the B+ tree are divided into hot leaf nodes and cold leaf nodes according to data access frequencies, wherein the hot leaf nodes are stored in the persistent memory PM, and the cold leaf nodes are stored in the ZNS SSD hard disk. The application utilizes the DRAM to quickly position the leaf nodes, reduces query time, stores the hot leaf nodes in the PM to meet real-time business read-write requirements, stores the hot data in the high-speed DRAM and the PM, reduces random read-write delay and performs well in big data analysis and other scenes.

The application provides a UCIe-based debugging information recording method, a storage medium and an artificial intelligence chip. The debugging information recording method comprises the following steps: setting the format of a debugging chain by a direct memory access module; updating the debugging chain by the direct memory access module; and sending the debugging chain to at least one of a memory bus and a control bus by the direct memory access module, wherein the debugging chain comprises a start mark, timestamp register information and an end mark, and further comprises at least one debugging register information or at least one error transmission information. The debugging information recording method, the computer readable storage medium and the artificial intelligence chip can effectively record debugging information.

The invention relates to an electronic system (100) comprising: - a controller module (3), with a core (31), peripheral (33) and memory (32) mapped on a memory bus (36), configured for executing instructions, - a supply module (1), with one or more power inputs (11) such as an energy harvester input, one or more power outputs (12) such as a load output, one or more ports (13) for connecting a storage element, configured for of supplying energy to different elements of the system, - a power and energy monitoring module (2), the power and energy monitoring module (2) being coupled to the controller module (3) by a first communication channel, the power and energy monitoring module (2) being coupled to the supply module 1) by a second communication channel. According to the invention, the supply module (1) is configured to collect information on the source, load and / or storage element, and the power and energy monitoring module (2) is configured for further processing the power and energy information received from the supply module and to allow the microcontroller module (3) to modify the order of execution of its instructions as a function of the information collected by the supply module (1).

The application discloses a communication auxiliary system, a master control device and a method, and belongs to the technical field of communication. The system adopts a mediation double-agent architecture, the master control device is logically connected between a communication terminal and a headset device, and meanwhile, isolated physical connections are established by simulating hands-free (HF) device specifications and audio gateway (AG) device specifications. The master control device is internally provided with an uplink data decision module and a memory bus with a first buffer address and a backup buffer address. When a local trigger event (such as touch, action or external entity button pulse) is captured, the decision module forcibly switches the memory data reading pointer of the digital audio by using the bottom-layer hardware atomic operation in an instant, and replaces the headset pickup source with an external high-definition pickup source. The whole switching process does not trigger the bottom-layer audio routing redistribution (Audio Policy) of the communication terminal operating system, and the absolute continuity of the Bluetooth bottom-layer SCO synchronous link frame sequence is maintained. The system realizes the high-response seamless migration and injection of the external pickup source without invading the protocol stack of the host device, supports the double-track solidification of the multi-dimensional voice shunting of the communication link, and significantly improves the communication and pickup efficiency in modern mobile and complex scenarios.

One embodiment provides circuitry coupled with cache memory and a memory interface, the circuitry to compress compute data at multiple cache line granularity, and a processing resource coupled with the memory interface and the cache memory. The processing resource is configured to perform a general-purpose compute operation on compute data associated with multiple cache lines of the cache memory. The circuitry is configured to compress the compute data before a write of the compute data via the memory interface to the memory bus, in association with a read of the compute data associated with the multiple cache lines via the memory interface, decompress the compute data, and provide the decompressed compute data to the processing resource.

To provide a memory device and a system that provide high speed access to a memory in higher temperature applications.SOLUTION: A method can include: receiving a plurality of consecutive commands on a unidirectional command-address (CA) bus input of a nonvolatile memory (NVM) device, the commands being synchronous with a timing clock; and for each received command, determining if the command is an express read (NVR) command, determining if a next consecutive command is a NVR command, in which consecutive NVR commands form a NVR command sequence; in response to the no more than the NVR command sequence, accessing read data stored in NVM cells; and driving the read data on parallel data input / outputs (I / Os) of the NVM device in a burst of data values, the data values of the burst being output in synchronism with rising and falling edges of the timing clock.SELECTED DRAWING: Figure 16

The application discloses a kind of virtual machine vcpu hanging death detection method under cloud computing scenario, belong to cloud computing field, and its specific include: in the back end of VMM, MMIO region is created to carry out memory mapping, by increasing configuration file, so that VMM is identified to the region, carries out I / O hardware abstraction, and registers device on memory bus using device tree mode, provides memory region to virtual machine, starts internal clock, calculates internal clock decrement time, and adjusts internal clock according to extracted guest time, uses misc device, starts high-precision timer and periodically sends heartbeat message, after sending, virtual machine enters VMM, and judges vcpu state by VMM, registers CPU hot plug mechanism, ensures that the vcpu of suspended state is offline and enters deadlock state, can quickly analyze virtual machine run dead problem and quickly locate fault.

A computing system including: a processing device; a main memory; a memory bus (e.g., a double date rate (DDR) bus) coupled between the processing device and the main memory; a random access memory; and a compute express link (CXL) fabric coupled between the processing device and the random access memory. The system is configured to identify a ratio, using a reinforcement learning technique, to manage distribution of memory usage across the main memory and the random access memory. For example, the system can measure performance metrics of the computing system as a result of the managing according to the ratio, and update, based on the performance metrics, a reward model configured to identify a ratio for a given state of the computing system to manage memory for maximized reward.

A system having: a compute express link fabric; a plurality of memory devices having random access memory cells, a memory bus; a main memory; at least one processing device connected to the main memory via the memory bus, and connected to the compute express link fabric; a peripheral bus; and a plurality of memory sub-systems connected to the at least one processing device via the peripheral bus. A plurality of portions of the random access memory cells in the plurality of memory devices are allocated respectively as a plurality of host memory buffers for the plurality of memory sub-systems. Each memory buffer can have random access memory cells allocated from more than one of the memory devices.

Systems, methods, and apparatus are described for accelerating access to borrowed memory via a network connection. For example, a memory management unit (MMU) of a computing device can be configured to connect to both random access memory via a memory bus and to a computer network via a communication device. The computing device can borrow an amount of memory from a remote device via a network connection using the communication device; and an application running in the computing device can use virtual memory addresses mapped to the borrowed memory. When a virtual address mapped to the borrowed memory is used, the MMU translates the virtual address to a physical address and instructs the communication device to access the borrowed memory.

Hardware-based training is described for synchronous dynamic random-access memory. In one aspect a memory includes a memory physical interface (PHY) of the memory coupled to a memory bus and configured to receive a received Pseudo-Random Binary Sequence (PRBS) through the memory bus at from a host PHY. A PRBS generator is configured to generate a generated PRBS and compare logic is configured to generate a score based on comparing the received PRBS to the generated PRBS. The memory PHY is further configured to send the score to the host PHY.

A method can include: receiving a plurality of consecutive commands on a unidirectional command-address (CA) bus input of a discrete nonvolatile memory (NVM) device, the commands being synchronous with a timing clock; for each received command, determining if the command is an express read (NVR) command, if a command is determined to be an NVR command, determining if a next consecutive command is an NVR command, wherein consecutive NVR commands form an NVR command sequence; in response to the no more than the NVR command sequence, accessing read data stored in NVM cells of the NVM device; and driving the read data on parallel data input / outputs (I / Os) of the NVM device in a burst of data values, the data values of the burst being output in synchronism with rising and falling edges of the timing clock; wherein the CA bus input includes a plurality of parallel CA signal inputs. Related memory devices and systems are also disclosed.

The invention provides a field bus data receiving and decoding method, and relates to the technical field of data processing. The method comprises the following steps: receiving petrochemical field bus data in real time, and further determining a lead code, a start delimiter, an end delimiter and effective data in sequence; the Manchester decoding module is used for performing Manchester decoding on valid data, sending the decoded data to the UART interface of the single chip microcomputer, performing Manchester encoding on the data received by the UART, and outputting the data to the field bus MAU interface through the SPI interface of the single chip microcomputer. According to the invention, the current situation that the development of the communication interface of the field bus transmitter needs to depend on a special field bus communication controller chip is solved, and meanwhile, two field buses, namely a foundation field bus and a PROFIBUS PA, are compatible. A hardware interface is simplified based on the unvarnished transmission function of the UART, a complex memory bus interface is not needed, user type selection of a single-chip microcomputer is facilitated, the development difficulty of a field bus transmitter can be reduced, and the cost and product miniaturization are further reduced.

The utility model relates to a USB (universal serial bus) transmission rate increasing circuit, belongs to the technical field of USB data transmission, and solves the problem of low transmission efficiency of a large number of file data generated by an infrared handheld thermal imager in the prior art. The USB transmission rate increasing circuit comprises a bus switching circuit and a memory. The bus switching circuit comprises a data selection circuit and a microcontroller; the data selection circuit comprises a low-speed external data interface, a high-speed external data interface and a synthetic data interface, and the synthetic data interface is connected with the microcontroller; the low-speed or high-speed external data interface is connected with the synthetic data interface; and the microcontroller is electrically connected with the memory. On the basis of keeping the external transmission USB2.0 interface of the original infrared handheld thermal imager, the transmission effect of externally exporting data at the USB3.0 speed is realized by adjusting the local circuit.

The application provides a SRAM type FPGA reconfiguration system with a self-refresh function based on a heterogeneous SIP and a working method, which comprises a heterogeneous SIP chip, a program memory NOR FLASH and a bus interface chip; the CPU, the SRAM type FPGA and the data memory DDR are integrated in the heterogeneous SIP chip; the CPU is used for refreshing the SRAM type FPGA, receiving reconfiguration instructions and reconfiguration data uploaded by a satellite-ground communication interface and reconfiguring the SRAM type FPGA; the SRAM type FPGA is used for on-orbit function reconfiguration; the data memory DDR is used for writing configuration bit streams of the FPGA; the program memory NOR FLASH is used for storing CPU programs and configuration bit streams of the FPGA; and the bus interface chip is used for transmitting the reconfiguration data uploaded by the satellite-ground communication interface to the CPU. The application can complete the refreshing and reconfiguration of the SRAM type FPGA in the SIP without additional external devices, reduces the influence of space single particles on the FPGA work, meets the space reliable application requirements, greatly reduces the cost, and reduces the volume, weight and power consumption.

A system having: a computer express link fabric; a plurality of memory devices having random access memory cells, a memory bus; a main memory; at least one processing device connected to the main memory via the memory bus, and connected to the computer express link fabric; a peripheral bus; and a plurality of memory sub-systems connected to the at least one processing device via the peripheral bus. A plurality of portions of the random access memory cells in the plurality of memory devices are allocated respectively as a plurality of host memory buffers for the plurality of memory sub-systems. Each memory buffer can have random access memory cells allocate from more than one of the memory devices.

Several examples of multi-port SRAM systems and related methods for distributed memory pools are disclosed. In one example, a system includes: a memory bus switch coupled to multiple memory buses; a memory pool switch coupled to one or more memory arrays; and one or more SRAM blocks coupled to the memory bus switch and the memory pool switch to convert a first transaction received on one of the multiple memory buses into a second transaction executed on the one or more memory arrays.

A memory controller (3) is provided with a read / write control circuit (11) for controlling a data read from a memory (5) and a data write to the memory (5), and a transfer control circuit (12) for controlling a data transfer in the inside of the memory (5). The transfer control circuit (12) transmits a first signal value that is unused in a JEDEC standard to the memory (5) via a signal line of a memory bus (4) and thereby enables a transfer command for writing the data read out from a source storage cell (61) to a destination storage cell (61), which is an additional transfer command not stipulated in the JEDEC standard, without going via the outside of the memory (5).

The invention relates to the field of image processing, and discloses a three-dimensional image processing system for laser scanning, which comprises a point cloud acquisition module, a coordinate mapping module, a storage alignment module and a feature extraction module, and is characterized in that the coordinate mapping module quantizes point cloud coordinates according to voxel scales and generates logic address tags; a storage alignment module calls a hardware cache line alignment width value as an address offset step length, and distributes physical storage pages for point clouds with different characteristics according to reflection intensity distribution; according to the invention, a pseudo shared path between sampling point data blocks is cut off by using an address isolation mechanism, so that a processor is ensured to maintain a continuous access state, the effective bandwidth utilization rate of a memory bus is effectively improved, and the memory performance is improved. And the real-time dynamic matching of the storage access capability and the hardware access demand is ensured.

The application is specifically an artificial intelligence-based system maintenance task intelligent scheduling system, and relates to the technical field of artificial intelligence, nonlinear dynamics and computer system operation and maintenance, comprising: a univariate time series data acquisition and denoising module; a phase space reconstruction module; a chaotic feature quantitative calculation module; a dynamics state discrimination and early warning module; and an entropy-driven task blocking and scheduling execution module.In the application, Takens embedding theorem is used to realize univariate micro time series signal acquisition, and the global dynamics state of the system can be completely characterized by only CPU interrupt response delay and memory bus bandwidth occupancy, thereby reducing data transmission, storage and calculation overhead; and the maximum Lyapunov exponent is calculated to realize accurate quantification of steady state, bifurcation and chaotic state.

The invention is applicable to the technical field of computers, and provides an image compression write direct memory access method, device and equipment and a storage medium, and the method comprises the following steps: receiving a to-be-compressed image containing a plurality of pixel components, and segmenting each row of each pixel component of the to-be-compressed image to obtain segmented to-be-compressed data; compressing all the segmented to-be-compressed data to obtain a segmented compressed code stream of each segmented to-be-compressed data; and obtaining a segmented DMA write address corresponding to the segmented compressed code stream, and writing the segmented compressed code stream into a memory corresponding to the segmented DMA write address through a memory bus, thereby realizing DMA write operation of to-be-compressed images of a plurality of pixel components through time division multiplexing of the same DMA controller by the plurality of pixel components. The complexity of the multi-pixel component image in DMA write operation is reduced, so that the area of a memory access chip can be saved, and the hardware cost is reduced.

A pseudo main memory system. The system includes a memory adapter circuit for performing memory augmentation using compression, deduplication, and / or error correction. The memory adapter circuit is connected to a memory, and employs the memory augmentation methods to increase the effective storage capacity of the memory. The memory adapter circuit is also connected to a memory bus and implements an NVDIMM-F or modified NVDIMM-F interface for connecting to the memory bus.

The application provides a real-time monitoring and root cause positioning method and system for software defects. The method collects real-time monitoring data of software processing threads, generates a thread competition heat map identifying shared resource contention points through time sequence correlation processing; triggers a server memory bus embedded hardware detection device based on the contention points to dynamically capture lock state transition signals, converts them into lock event sequences; sorts and constructs a lock contention trajectory map according to timestamps, and completely displays the process path of the lock mechanism from acquisition to release; constructs a thread and lock dependency topology based on the trajectory map, identifies resource contention chains, maps the contention chains into unsynchronized shared variable access paths in the source code, and triggers real-time alarm signals. The application accurately locates the unsynchronized shared variable access path, improves the lock competition diagnosis efficiency, and reduces the risk of system deadlock to zero.

The application discloses a convolution accelerator based on RISC-V instruction set architecture expansion, which is located in a computer system and is physically connected with a memory through a memory bus, introduces a data buffer, and uses a buffer block strategy to improve convolution operation data multiplexing and reduce the demand of convolution operation on memory access, wherein the modules in the convolution accelerator include a fetch and instruction distribution module, a data carrying module, a feature map buffer, a convolution kernel buffer, a convolution operation unit, a convolution logic operation control register group, an output buffer and an adder; and the convolution accelerator is provided with a self-defined expansion instruction. The application also discloses a method for accelerating convolution operation by using the convolution accelerator.

This invention relates to fused processing of continuous mathematical operators. Systems and methods for fused processing of continuous mathematical operators are disclosed. Fused processing of continuous mathematical operations (e.g., pointwise nonlinear functions) without storing intermediate results to memory improves performance when memory bus bandwidth is limited. In one embodiment, continuous mathematical operations including at least two of convolution, upsampling, pointwise nonlinear function, and downsampling are performed to process input data and generate aliasing-free output data. In one embodiment, input data is spatially blocked for parallel processing such that intermediate results generated during processing of each block of input data can be stored in shared memory within the processor. Storing intermediate data in shared memory improves performance compared to storing and loading intermediate data to and from external memory.

An accelerator system includes an off-chip memory, a bus, and an accelerator. The off-chip memory includes a first buffer. The accelerator includes a second buffer, a third buffer, a fourth buffer, an instruction fetching module, a load module, a storage module, a convolution control module, and a computation module. The instruction fetching module fetches instructions from the off-chip memory via the bus. The load module loads input data from the off-chip memory via the bus. The storage module stores the output data to the off-chip memory via the bus. The computation module performs convolution computations based on the input map and multiple kernels allocated by the convolution module. In response to loading the input data or instruction from the off-chip memory, the requested input data or instruction is transferred from the off-chip memory and temporarily stored in the first buffer.

A memory subsystem includes link encryption for the system memory data bus. The memory controller can provide encryption for data at rest and link protection. The memory controller can optionally provide link encryption. Thus, the system can provide link protection for the data in transit. The memory module can include a link decryption engine that can decrypt link encryption if it is used, and performs a link integrity check with a link integrity tag associated with the link protection. The memory devices can then store the encrypted protected data and ECC data from the link decryption engine after link protection verification.