Calculation optimization method and device of shallow depth model based on systolic array

Abstract

The invention provides a calculation optimization method and device for a shallow depth model based on a systolic array. By determining the standard input depth / standard output depth of the systolic array and the input depth and the output depth of the shallow depth model, determining the input splicing number and the output splicing number. And according to the input splicing number and the output splicing number value, splicing the input data and / or the corresponding output data so as to obtain the input data in batches and provide the input data to the systolic array in parallel, and / or caching a plurality of output data corresponding to the input data through an output buffer and then outputting the output data in batches, therefore, the migration efficiency of the input data and / or the output data can be improved, and the utilization efficiency of the systolic array can also be improved.

Description

technical field [0001] The present disclosure relates to the field of computers, and more particularly, to a calculation optimization method and device based on a shallow depth model of a systolic array. Background technique [0002] In the field of data processing, many data processing processes, such as deep learning processes, can be processed through convolutional neural networks (CNN), which require a large number of calculation steps, most of which are multiply-accumulate operations. [0003] Systolic array structures are currently used to implement a large number of multiply-accumulate operations. The systolic array structure includes computing units set at the nodes of each row and column (denoted by PE, each PE includes a multiplication accumulator or similar computing device, or a storage unit), the advantage lies in the operands involved in internal operations or intermediate The result can be performed in a high-speed parallel pipeline, that is, each piece of in...

AI Technical Summary

Problems solved by technology

[0004] In the calculation process based on the systolic array structure, the utilization efficiency of the systolic array structure, the transfer efficiency of moving the input data to the input buffer and taking the output data calculated by the systolic array from the output buffer corresponding to the systolic array are in the Factors that are often considered in the calculation process, and in the case of calculation processes applied to shallow depth models, the utilization efficiency of the systolic array structure and the transfer efficiency of data are often not high

Images