Computing system

Abstract

Provided is a computing system capable of obtaining a calculation speed exceeding that of 16-bit floating point processing while maintaining accuracy of calculation results. A computing system 1 includes a parameter server 10, a communication path 20 and a worker 30. The parameter server has a storage unit 11 that stores a parameter value of a training target model represented by a 32-bit floating point number, and a first conversion unit 12 that converts the parameter value into data represented by a floating point number with 10 bits or less, in which a mantissa is 3 bits or less and an exponent is 6 bits or less. The communication path transmits the data transmitted and received between the parameter server and the worker. The worker has a processing unit 31 that computes a product and a sum of the data. The parameter server further has a second conversion unit 13 that converts the data received from the worker into an updating difference represented by a 32-bit floating point number, and an updating unit 14 that updates the parameter value on the basis of the updating difference.

Dependencies are inferred from the claim text.