Method and apparatus for vector process facing sliding window

Abstract

The present invention relates to high efficiency processor designs, in the field of digital signal processing and intelligent computing, a method for slide window vector processing is disclosed, includes a set of vector instructions, a vector instruction designates a first source vector to perform corresponding operation with a second source vector according to an operation type, a sliding windowoperation skip step number is specified by a value in a third scalar register, and result elements obtained by the operation are sequentially written into a result vector register. The invention alsodiscloses a device for sliding window vector processing. A method and apparatus for processing a sliding window vector, The traditional vector and windowed vector operations are uniformly represented,and the vector processing is realized by a unified vector instruction and control method, which improves the performance of the vector processing device, and can flexibly represent various sliding window vector operations, and meets the needs of many new tasks with window operations.

Description

technical field [0001] The invention relates to the fields of high-efficiency processor design, digital signal processing and intelligent computing, in particular to a method and device for sliding window vector processing. Background technique [0002] Computation-intensive applications such as machine learning and digital signal processing have greatly increased the demand for applications driven by the rapid development of informatization. The use of vector processing methods to support batch processing of large amounts of data plays an important role in these applications. critical to device performance. [0003] A vector is a set of data stored in a computing device, such as Figure 4 As shown, in the processing, a vector is composed of N elements (N is a positive integer greater than or equal to 1), if the vector is stored continuously, its storage location is specified by the vector head address and the length of the vector, if the vector is not stored continuously, t...

AI Technical Summary

Problems solved by technology

But now the traditional vector processing device is generally used to process two source vectors of equal length, and each element in the vector performs the same operation (such as addition, subtraction, multiplication, etc.), and the operation with window processing characteristics is in the traditional vector processing device In general, it is necessary to expand the window vector and source vector into vectors that can directly calculate the corresponding elements, so that traditional vector instructions can be used for processing, which increases the demand for storage space and increases the amount of calculation for data conversion.

Images