39 results about "Matrix compression" patented technology

The invention belongs to the technical field of GNSS-R SAR (Global Navigation Satellite System-Radar Synthetic Aperture Radar) super-resolution imaging, and discloses a fast distance super-resolution imaging method based on a GNSS-R SAR. According to the method, matrix compression, regularization modeling and a rapid optimization algorithm are creatively combined, and a set of efficient and stable distance super-resolution processing flow is formed. Specifically, after a preliminary imaging result of a GNSS echo signal is obtained and a signal convolution model form of the GNSS echo signal is given, firstly, dimension reduction processing is performed on an observation matrix in a signal convolution model through singular value decomposition, and on the basis of a weighted singular value maintenance strategy processing result, the signal convolution model is reconstructed by using an inverse matrix of a truncated measurement matrix; then, starting from a regularization strategy, introducing an L1 norm constraint to construct a target function by utilizing the sparse characteristic of a target; and finally, solving the target function by adopting a rapid iterative optimization algorithm. According to the method provided by the invention, the range resolution of the GNSS echo data is remarkably improved.

The present invention belongs to the technical field of GNSS-R SAR super-resolution imaging, and discloses a fast range super-resolution imaging method based on GNSS-R SAR. The present invention innovatively combines matrix compression, regularization modeling and fast optimization algorithm to form a set of efficient and stable range super-resolution processing procedures. Specifically, after obtaining the preliminary imaging results of the GNSS echo signal and giving its signal convolution model form, the observation matrix in the signal convolution model is first reduced in dimension by singular value decomposition, and on the basis of the processing results of the weighted singular value preservation strategy, the signal convolution model is reconstructed using the inverse matrix of the truncated measurement matrix; then, starting from the regularization strategy, the L1 norm constraint is introduced to construct the objective function using the sparse characteristics of the target; finally, a fast iterative optimization algorithm is used to solve the objective function. The method of the present invention significantly improves the range resolution of GNSS echo data.

The invention discloses a multi-satellite joint task planning method and system for remote sensing observation, and relates to the field of remote sensing satellite task planning, and the method comprises the steps: carrying out the time window slice discretization of obtained satellite task data according to a matrix compression coding method, and obtaining a compression matrix; traversing and scanning each element in the compression matrix, removing invalid observation windows according to the elements, and aggregating the remaining valid time windows by using the time slice decision matrix to obtain the startup state of each satellite in each valid coverage time slice; in multi-satellite joint task planning, grid point division is carried out on a target area, and the joint coverage rate of multiple satellites in an overlapped time window is calculated; performing explosion decision on the time slice decision matrix through a multi-target fireworks algorithm to obtain a multi-satellite collaborative observation planning scheme; the multi-satellite joint task planning method and system aim at solving the problems that remote sensing satellites are low in scheduling efficiency, uneven in resource utilization and complex in coverage rate calculation in a multi-task and multi-constraint scene.

The disclosed embodiments relate to matrix compression / decompression instructions. In one example, a processor includes fetch circuitry to fetch a compression instruction having a format with fields to specify positions of an opcode and a decompressed source matrix and a compressed destination matrix; the decoding circuit is used for decoding the extracted compression instruction; and execution circuitry to, in response to the decoded compression instruction, generate a compressed result according to a compression algorithm by compressing a specified decompressed source matrix via any of: compress non-zero-value elements together and store a matrix position of each non-zero-value element in a header, or using fewer bits to represent one or more elements and using a header to identify matrix elements represented by fewer bits; and storing the compressed result to the specified compressed destination matrix.

The invention discloses a tensor autoregression model coefficient estimation optimization system and method based on a sketch matrix, and relates to the technical field of multi-dimensional time series data processing. The system comprises a multi-modal data acquisition module, a self-adaptive sketch matrix construction module, a tensor autoregression model construction module, a multi-stage coefficient estimation optimization module and the like. The method comprises the following steps: preprocessing multi-source time sequence data to construct a tensor, generating sparse sketch matrix compressed data based on PARAFAC decomposition and differential sampling, constructing a tensor autoregression model containing a lag term, optimizing a coefficient by combining an L1-L2 mixed regular term and an ADMM algorithm, dynamically updating the model, and deploying the model. According to the method, the calculation complexity is reduced through the sketch matrix, the comprehensive efficiency is improved by more than or equal to 50%, data key information retention is guaranteed, the reconstruction error is less than or equal to 5%, the prediction precision is reduced by less than or equal to 10%, edge equipment is adapted, the method can be applied to scenes such as Internet of Things monitoring and financial analysis, the contradiction between efficiency and precision in high-dimensional data processing is solved, and the operation and maintenance cost is reduced.

Disclosed are systems and techniques for performing matrix multiply operations on an expressive sparse matrix representation with limited metadata. The techniques include receiving a sparse matrix, metadata corresponding to the sparse matrix, and a matrix operand. The sparse matrix contains a first number (N) of elements to retain from a dense matrix which comprises at least a second number (M) of elements. The metadata corresponding to the sparse matrix is based on a third number (P) of positions and a format determined during compression of the dense matrix. The techniques include selecting, by one or more selection circuits, a subset of elements of the matrix operand based on the metadata corresponding to the sparse matrix and performing one or more matrix multiply operations on the sparse matrix and the subset of elements of the matrix operand.

The embodiment of the invention provides a data processing method and device, electronic equipment and a storage medium. The data processing method comprises the following steps: acquiring a first object matrix and a second object matrix which are used as matrix multiplication multipliers; compressing the first object matrix into a first target matrix and compressing the second object matrix into a second target matrix according to the first compression instruction; and multiplying the first target matrix and the second target matrix according to the first matrix multiplication operation instruction to obtain a first operation result matrix, and obtaining a matrix multiplication operation result of the first object matrix and the second object matrix according to the first operation result matrix. According to the method, the matrix compression and the matrix operation can be separated and executed in parallel, so that the flexibility of the matrix compression is improved, the design of a matrix operation instruction is simplified, and the matrix operation time is greatly shortened.

This invention discloses a parallelizable adaptive lossless matrix compression method. It involves dividing a large first matrix into smaller second matrices, assigning a splitting process to each second matrix, and having these splitting processes divide all second matrices into multiple bit matrices in parallel. Compression threads are then assigned to each bit matrix, and these compression threads adaptively select the optimal compression strategy from a set of compression strategies. Finally, the compression threads compress all bit matrices in parallel according to the determined compression strategy, forming a compressed matrix and a flag array as the final compressed data of the first matrix. By hierarchically and parallelizing the compression process, the memory overhead of the compression process is effectively reduced and the compression speed is improved. Adaptive compression is achieved by selecting the optimal compression strategy through trial compression, effectively improving the data compression ratio. Lossless data compression is achieved by constructing a flag array.

The invention discloses an MIMO wireless communication channel matrix compression calculation method and device, and the method comprises the steps: marking the position of a non-zero element in a channel matrix as 1, marking the position of a zero element as 0, and generating a binary index matrix; performing addressing and aggregation operation on non-zero elements in the channel matrix according to a column priority based on the binary index matrix, and sequentially writing the non-zero elements into a preset continuous storage space in the memory according to the column priority to generate a compression matrix; and sequentially aiming at each calculation unit, firstly determining a corresponding index column according to the binary index matrix, counting non-zero elements in the index column, if the count is greater than 0, performing multiplication and addition operation based on a multiplication and addition operation module, and if the count is equal to 0, writing a 0 value into a partial sum accumulation register as an operation result of the current calculation unit. According to the scheme of the invention, the storage overhead of the MIMO wireless communication channel matrix can be reduced, and the calculation efficiency of the channel matrix can be improved.

The present application relates to the technical field of fault diagnosis, in particular to a power grid fault analysis method and system, which comprises constructing a fusion tensor and performing dimension reduction processing on the fusion tensor to generate a fusion feature matrix; constructing a convolutional autoencoder model to detect abnormal fusion feature matrix; classifying power grid faults according to abnormal data of the abnormal feature matrix, and calculating power grid equipment fault area based on the result of power grid fault classification. The present application has the beneficial effect of ensuring the integrity and accuracy of data by using a high-precision low-rank tensor completion algorithm to complete power grid operation data. Key features are extracted from multi-dimensional data collected by multiple source sensors and a fusion feature matrix is constructed using Tucker decomposition and matrix compression technology, effectively improving the efficiency of data processing and the ability of feature expression. The application of the convolutional autoencoder model enhances the detection accuracy of abnormal features, and the improved weighted summation and hierarchical clustering-based method improves the accuracy of fault classification and positioning.

A system and method for performing multiplication for a neural network, e.g. for data of one or more layers in a neural network, may include loading a portion of a compressed version of a sparse input matrix into a cache memory; uncompressing a subset of the data in the portion of the compressed version of the sparse input matrix; and multiplying a sparse kernel matrix by the subset of the data using a set of instructions which are themselves created based on the sparse kernel matrix.

The invention discloses a multi-mode fretting wear calculation method, device and system and a storage medium. The multi-mode fretting wear calculation method comprises the steps that S1, dynamic node data is obtained, and the wear amount is calculated; s2, calculating a geometric self-adaptive wear direction; s3, performing node classification and index optimization and adaptive adjustment of an increment step; and S4, carrying out million-level node parallel computing by adopting a block storage and matrix compression technology, and controlling the grid scanning times through KMESHSWEEP parameters. By adopting the technical scheme of the invention, the technical problems of poor geometric adaptability, low algorithm convergence and limited calculation scale in fretting wear simulation are solved.

The invention discloses a fully expressive sparse matrix representation with finite element data. Systems and techniques for compressing a dense matrix into a fully expressive sparse matrix representation with finite element data are disclosed. The techniques include generating a sparse matrix with corresponding metadata based on the dense matrix. Generating a sparse matrix with respective metadata includes: identifying a first number (M) of elements to be compressed, a second number (N) of elements to be retained, and a third number (B) of bits used by each metadata value; determining a metadata value of each of the N elements of the dense matrix; packing the first metadata value having more than B bits into a second metadata value having B bits; and generating a sparse matrix containing the N elements of the dense matrix. The techniques include storing a sparse matrix and corresponding metadata, wherein the corresponding metadata includes a second metadata value.

The invention relates to a method for testing interface shear strength of an optical fiber embedded composite material based on an annular pressure head, and belongs to the technical field of interface performance testing of optical fibers and composite materials. Comprising the following steps: providing a group A sample and a group B sample of which the cross sectional areas are smaller than the area of a test pressure head; the A group of samples are composite material samples embedded with optical fibers, the end surfaces of the optical fibers and the surface of the matrix are coplanar, and the B group of samples are composite material samples not embedded with optical fibers; acquiring annular pressure data; obtaining matrix data; calculating shear strength; correcting the matrix compression load based on the ratio of the effective compression area of the A group of samples to the effective compression area of the B group of samples to obtain a matrix equivalent compression load; and calculating the difference between the total load and the equivalent compression load of the matrix to obtain an interface shear force, and further calculating the interface shear strength. The interference of the plastic deformation of the matrix on a test result is eliminated, and the accuracy and the reliability of a calculation result are remarkably improved.

The invention provides a high-order sparse matrix LDL efficient decomposition calculation acceleration method, which can be used for a high-order sparse matrix linear equation set solving calculation scene. The accelerator is constructed based on an FPGA platform, and efficient decomposition of a high-order sparse matrix is achieved through a software and hardware cooperative computing mechanism. The software level realizes symbolic analysis of a sparse positive definite matrix, calculation sequence dependency modeling and LDL decomposition hardware customization instruction generation; and the hardware acceleration unit realizes streamlined execution of decomposition tasks by performing instruction fetching, decoding and execution of customized instructions, and completes high-order matrix numerical calculation. The system comprises an instruction generation unit, an instruction analysis unit, a calculation module, a data access module and a control scheduling module, and all the modules work cooperatively to improve the degree of parallelism and the storage access efficiency. Matrix data types support fp16, fp32 and fp64, matrix compression is realized by adopting a sparse column (CSC) format and a matrix rearrangement method, and storage overhead and memory access delay are remarkably reduced while high calculation precision is ensured. Experiments show that compared with traditional CPU and GPU platforms, the sparse matrix solution performance and energy efficiency of the method are remarkably improved by one order of magnitude, and the method is suitable for high-precision engineering calculation and edge real-time solution scenes such as engineering simulation, structural mechanics, electromagnetic analysis and attitude calculation.

The invention relates to a weight matrix compression method, device and equipment for a high-computing-power chip and a storage medium product. The method comprises the following steps: performing bit slicing processing on an original weight matrix to obtain at least two bit slicing matrixes; classifying the at least two bit slice matrixes to obtain at least one first bit slice matrix and at least one second bit slice matrix; performing coding processing on the at least one second bit slice matrix to obtain at least one coded second bit slice matrix; determining a compressed weight matrix according to the at least one first bit slice matrix and the at least one coded second bit slice matrix; the sparse rate of the second bit slice matrix is greater than the sparse rate of the first bit slice; and the compressed weight matrix is applied to the model training of the Transform network. By adopting the method, the weight matrix storage load can be reduced, so that the memory access speed is accelerated, and the overall performance of the system is improved.

The present application discloses a key-value compression method, a large language model, and an electronic device in a self-attention mechanism, and relates to the field of computer technology. The compression method includes performing multiple residual decompositions on the key matrix and the value matrix, respectively, to obtain a key residual vector and a value residual vector after each decomposition; clustering and compressing the key residual vector and the value residual vector after each decomposition, and performing attention calculation on the query matrix, the compressed key residual vector, and the value residual vector; and accumulating all attention calculation results. The present application solves the problem that the Linear Transformer cannot use the standard Softmax Transformer parameters and has significant differences from the standard Softmax Transformer.

The invention provides an efficient solution calculation acceleration method for a high-order sparse matrix linear equation set. The high-order sparse matrix linear equation set is widely applied to the fields of scientific calculation, engineering simulation and the like. A traditional FPGA-based solving method is low in hardware resource utilization rate and complex in data flow scheduling, so that the calculation efficiency is limited, and the efficient solving requirement of a large-scale sparse matrix is difficult to meet. In order to solve the problem, the invention provides a high-order sparse matrix-oriented linear equation set efficient solution calculation acceleration method, a sparse matrix compression storage format is combined, a linear equation set solution special circuit and an acceleration solution algorithm based on LDL decomposition are designed, and the calculation efficiency is effectively improved.

Systems and techniques are described herein for compressing data used in cryptographic operations. For example, a process may include obtaining a first data structure, wherein the first data structure comprises polynomials; generating a second data structure based on the first data structure, wherein the second data structure comprises coefficients of the polynomials; sorting the second data structure in an ascending order to obtain a sorted second data structure; updating the sorted second data structure based on differences between elements of the sorted second data structure to obtain a delta-encoded data structure; performing an entropy coding on the delta-encoded data structure to obtain an entropy-encoded output; recovering an updated first data structure using the entropy-encoded output, wherein the updated first data structure corresponds to the first data structure with a different order of first data structure elements; and performing a cryptographic operation using the updated first data structure.

The utility model relates to the technical field of compression molding device, and disclose a kind of matrix compression molding device, and the anti-skid bottom plate is fixedly sleeved in the position around the surface of fixed bottom plate, first cross plate is movably connected in the position of both sides of fixed bottom plate top, second cross plate is movably connected in the position of both sides of first cross plate on fixed bottom plate top, second screw rod is movably connected in the position of the bottom of first cross plate in fixed bottom plate interior, first screw rod is movably connected in the position of the bottom of second cross plate in fixed bottom plate interior, servo motor is drivingly connected in the position of both ends of second screw rod and first screw rod, the utility model is equipped with servo motor, first screw rod, second screw rod, first cross plate, second cross plate, is favorable to after matrix compression molding, by servo motor drive first screw rod, second screw rod is rotated, and then make first cross plate, second cross plate separate to around, and then quickly the matrix of compression molding is demoulded, increase the production efficiency of matrix.

Embodiments herein relate to a neural network compression technique in which a weight matrix within a neural network is transformed via matrix multiplication with an orthogonal matrix. The orthogonal matrix is derived from a calibration data set, which is generally selected to widely represent expected runtime input data, and the transformation causes the resulting modified weight matrix to have components ordered by relative importance. The modified weight matrix is incorporated into a compressed neural network with fewer weights. By removing one or more components of lower importance, the size of the neural network (and thus its storage and execution overhead) is reduced while still maintaining acceptable performance levels.

The present invention provides a method and device for compressed storage of a finite element assembly matrix for process simulation. The method for compressed storage of a finite element assembly matrix for process simulation includes: using two arrays to store the positions and values of the elements of the matrix to be stored respectively. By using two arrays to store the positions and values of the elements of the matrix to be stored respectively, this method requires less space compared to directly storing the matrix to be stored, can greatly reduce the storage overhead required for the matrix to be stored, improve the memory usage efficiency in the numerical calculation process of the corresponding process simulation, and enhance the ability of the corresponding process simulation software to handle large-scale meshes. Moreover, this method has simple steps, is easy to implement, the operation process is transparent to users, does not affect the accuracy of the numerical calculation method for process simulation, does not significantly affect the efficiency of the numerical calculation in the process simulation process, improves the memory space usage efficiency, enhances the ability of the software to handle large-scale meshes, and has broad application prospects.

Systems and methods for decompressing a meaningful sparse matrix representation with limited metadata are disclosed. The methods include receiving a sparse matrix and metadata corresponding to the sparse matrix. The sparse matrix is ​​a compressed representation of a dense matrix. The sparse matrix contains a first number N of elements to be retained from the dense matrix, which comprises at least a second number M of elements. The metadata corresponding to the sparse matrix is ​​based on a third number P of positions and a format determined during the compression of the dense matrix. The methods include generating an uncompressed matrix based on the sparse matrix and the metadata corresponding to the sparse matrix.

Provided is a new coupling matrix compression device for compressing an input coupling matrix corresponding to an input Ising Hamiltonian in an optimal solution search device that compresses the input coupling matrix, solves the input coupling matrix with an Ising machine using a coupling matrix obtained through compression, and restores a solution corresponding to the input coupling matrix from an output solution of the Ising machine. A coupling matrix compression device according to the disclosed technology comprises a compression determination unit. The compression determination unit reduces the comparison between the value of an Ising Hamiltonian in cases where the spins of the spin pair of interest have the same sign and the value of the Ising Hamiltonian in cases where the spins of the spin pair of interest have different signs to a number division problem and carries out evaluation, thereby determining the possibility of compression of the spin pair of interest, where, among spins coupled by an input coupling matrix or by a combined matrix after compression, the spin pair of interest is a spin pair to be subjected to compression determination, and peripheral spins are spins excluding the spin pair of interest.

Systems and methods for performing matrix multiplication operations with a meaningful sparse matrix representation containing limited metadata are disclosed. The methods include receiving a sparse matrix, metadata corresponding to the sparse matrix, and a matrix operand. The sparse matrix contains a first number N of elements, which are to be retained from a dense matrix comprising at least a second number M of elements. The metadata corresponding to the sparse matrix is ​​based on a third number P of positions and a format determined during compression of the dense matrix.The procedures involve selecting a subset of elements of the matrix operand by one or more selection circuits based on the metadata corresponding to the sparse matrix, and performing one or more matrix multiplication operations on the sparse matrix and the subset of elements of the matrix operand.

Systems and methods for compressing a dense matrix into a meaningful sparse matrix representation with limited metadata are disclosed. The methods include generating a sparse matrix with corresponding metadata based on a dense matrix. Generating the sparse matrix with corresponding metadata involves identifying a first number M of elements to be compressed, a second number N of elements to be retained, a third number P of positions, and a format; determining a metadata value for each of the N elements of the dense matrix based on the identified P and the identified format, wherein the dense matrix has at least M elements; and generating the sparse matrix containing the N elements of the dense matrix.The methods involve storing the sparse matrix and the corresponding metadata, where the corresponding metadata includes the metadata value for each of the N elements of the dense matrix.

Disclosed are systems and techniques for compressing a dense matrix into an expressive sparse matrix representation with limited metadata. The techniques include generating a sparse matrix with corresponding metadata based on a dense matrix. Generating the sparse matrix with corresponding metadata includes identifying a first number (M) of elements to compress, a second number (N) of elements to retain, a third number (P) of positions, and a format; determining a metadata value for each of N elements of the dense matrix based on the identified P and the identified format, wherein the dense matrix includes at least M elements; and generating the sparse matrix containing the N elements of the dense matrix. The techniques include storing the sparse matrix and the corresponding metadata, wherein the corresponding metadata comprises the metadata value for each of the N elements of the dense matrix.

The invention discloses an MIMO (Multiple Input Multiple Output) wireless communication channel matrix compression method and device. The method comprises the following steps: dividing a channel matrix into a plurality of matrix sub-blocks with the same size; determining a sparse mode index value of each matrix sub-block according to the type of each matrix sub-block; storing each sparse mode index value into an index addressing array of a memory according to the numbering sequence of each matrix sub-block; for each sparse sub-block, generating a respective index column of each sparse sub-block, and storing each index column into an index addressing array according to the numbering sequence of each sparse sub-block; when matrix data are stored, all element data are stored in a block sparse channel array of a memory for dense sub-blocks, only all non-zero element data are stored in the block sparse channel array for sparse sub-blocks, and data are not stored for all-zero sub-blocks. According to the invention, the technical effect of effectively reducing the storage overhead of the channel matrix in MIMO wireless communication is realized.

The embodiment of the invention provides a data accumulation method based on activation value matrix compression and a near storage computing system, and the method comprises the steps: obtaining an activation value matrix, carrying out the region division of the activation value matrix, and obtaining a to-be-compressed region; compressing the to-be-compressed region based on the numerical value of the to-be-compressed region to obtain a compression activation matrix; and performing operation on the compression activation matrix based on a preset weight matrix to obtain a data accumulation result. According to the invention, the problem of high data handling energy consumption of the near-storage computing system in the prior art is solved, and the effect of reducing the data handling energy consumption of the near-storage computing system is achieved.

This invention discloses a hierarchical TCAM message parsing and field extraction system based on parse graph optimization, relating to the field of communication technology. It includes: a software calculation and configuration module for performing graph topology optimization on the protocol parsing graph to generate multiple independent parsing subgraphs; generating a hardware resource configuration table corresponding to each parsing subgraph based on the key bit set corresponding to each parsing subgraph, and distributing the hardware resource configuration table to the key bit extraction unit, mapping table matching unit, flow classification module, and field extraction module on the hardware side, thereby configuring the logic structure of the target circuit according to the hardware configuration information and realizing dynamic reconfiguration of the protocol parsing function. This invention decomposes the protocol parsing graph into parsing subgraphs and adopts a multi-level unit fusion architecture at the hardware level to achieve pipelined protocol parsing. Combined with key bit matrix compression technology, it can significantly reduce TCAM storage requirements and effectively solve the timing and storage bottlenecks in large-scale message header parsing.