Computing engine, data processing method thereof, processing chip and computer equipment

By designing dot product unit groups that support different precisions in the computing engine, the problem of flexibility and efficiency in low-precision data computing in large artificial intelligence models is solved, achieving efficient multi-precision data processing and reducing computing costs.

CN121785558APending Publication Date: 2026-04-03MOORE THREADS TECH CO LTD
0 Cites 0 Cited by

Patent Information

Application Number
CN202511948922.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-12-22
Publication Date
2026-04-03

AI Technical Summary

Technical Problem

Existing technologies struggle to efficiently support computations of low-precision data formats in large AI models, especially in mixed-precision training models, where flexibly handling data of varying precision becomes a challenge.

Method used

Design a computation engine in which each dot product unit includes a first, second, and third computation unit group that supports dot product operations of data with different precisions. Data can be flexibly transmitted through a control logic unit to realize the operation of low-precision data and support the computational needs of multiple precisions.

Benefits of technology

It improves flexibility and efficiency in data calculations with varying precision, reduces computational costs, and enhances the adaptability and speed of the computing engine.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention provides a calculation engine and a data processing method thereof, a processing chip and computer equipment, the calculation engine comprises a dot multiplication unit array and a control logic unit, the dot multiplication unit array comprises a plurality of dot multiplication units, and each dot multiplication unit comprises a first calculation unit group and a second calculation unit group, the point multiplication unit supports point multiplication of data equal to or lower than the first precision; a second calculation unit group supporting point multiplication of data equal to or lower than a second precision, the second precision being lower than the first precision; the third calculation unit group supports point multiplication operation of data with third precision, the third precision is lower than the second precision, the control logic unit reads data to be processed according to an obtained control instruction and transmits the data to the point multiplication unit array, and when the data to be processed is the data with the third precision, the point multiplication operation of the data to be processed is carried out. And the control logic unit transmits the data with the third precision to the dot multiplication unit according to a data processing mode.
Need to check novelty before this filing date? Find Prior Art