公开(公告)号：US20230161835A1

公开(公告)日：2023-05-25

申请号：US18093929

申请日：2023-01-06

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Tao LI ,  Tingyu LU ,  Baolong CUI ,  Licheng YU ,  Haocheng LIU ,  Weibin LIN

IPC: G06F17/16 ,  G06F13/28

CPC classification number: G06F17/16 ,  G06F13/28 ,  G06F2213/28

Abstract: A matrix operation method is provided, applied to an accelerator configured to perform a matrix operation. A matrix operation accelerator (100) respectively stores, in response to a received matrix operation instruction, subsets of a first matrix and subsets of a second matrix in a first storage space and a second storage space of a memory (120); stores, in a third storage space of the memory (120), subsets obtained after the subsets of the first matrix are multiplied by the subsets of the second matrix; and performs matrix operations on the subsets of the first matrix and the subsets of the second matrix based on the matrix operation instruction, to obtain matrix operation results. The dedicated matrix operation accelerator (100) is used to perform a matrix operation, so that a large-scale matrix operation can be completed in relatively short time, thereby offloading a matrix operation burden of a processor.

公开(公告)号：US20210294623A1

公开(公告)日：2021-09-23

申请号：US17338137

申请日：2021-06-03

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Haocheng LIU ,  Shiping DENG

IPC: G06F9/455 ,  G06F9/50 ,  G06F12/02 ,  G06F9/54

Abstract: This application provides a data processing method. The method is applied to a server, the server includes a first container and an Android driver (240), and a first service manager is deployed in the first container. Each container is allocated its respective private memory (2412, 2422, and 2432) and private service linked list (2411, 2421, and 2431), so as to avoid a problem in a conventional technology of data inconsistency and affected service processing caused by sharing of same memory by a plurality of containers.