Processing

Please wait...

PATENTSCOPE will be unavailable a few hours for maintenance reason on Tuesday 26.10.2021 at 12:00 PM CEST
Settings

Settings

Goto Application

1. WO2019015541 - CALCULATION METHOD AND RELATED PRODUCT

Publication Number WO/2019/015541
Publication Date 24.01.2019
International Application No. PCT/CN2018/095706
International Filing Date 13.07.2018
IPC
G06F 9/30 2006.1
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
FELECTRIC DIGITAL DATA PROCESSING
9Arrangements for program control, e.g. control units
06using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
30Arrangements for executing machine instructions, e.g. instruction decode
CPC
G06F 12/023
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
FELECTRIC DIGITAL DATA PROCESSING
12Accessing, addressing or allocating within memory systems or architectures
02Addressing or allocation; Relocation
0223User address space allocation, e.g. contiguous or non contiguous base addressing
023Free address space management
G06F 7/50
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
FELECTRIC DIGITAL DATA PROCESSING
7Methods or arrangements for processing data by operating upon the order or content of the data handled
38Methods or arrangements for performing computations using exclusively denominational number representation, e.g. using binary, ternary, decimal representation
48using non-contact-making devices, e.g. tube, solid state device; using unspecified devices
50Adding; Subtracting
G06F 7/523
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
FELECTRIC DIGITAL DATA PROCESSING
7Methods or arrangements for processing data by operating upon the order or content of the data handled
38Methods or arrangements for performing computations using exclusively denominational number representation, e.g. using binary, ternary, decimal representation
48using non-contact-making devices, e.g. tube, solid state device; using unspecified devices
52Multiplying; Dividing
523Multiplying only
G06F 9/3001
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
FELECTRIC DIGITAL DATA PROCESSING
9Arrangements for program control, e.g. control units
06using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
30Arrangements for executing machine instructions, e.g. instruction decode
30003Arrangements for executing specific machine instructions
30007to perform operations on data operands
3001Arithmetic instructions
G06F 9/30036
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
FELECTRIC DIGITAL DATA PROCESSING
9Arrangements for program control, e.g. control units
06using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
30Arrangements for executing machine instructions, e.g. instruction decode
30003Arrangements for executing specific machine instructions
30007to perform operations on data operands
30036Instructions to perform operations on packed data, e.g. vector operations
G06F 9/30043
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
FELECTRIC DIGITAL DATA PROCESSING
9Arrangements for program control, e.g. control units
06using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
30Arrangements for executing machine instructions, e.g. instruction decode
30003Arrangements for executing specific machine instructions
3004to perform operations on memory
30043LOAD or STORE instructions; Clear instruction
Applicants
  • 上海寒武纪信息科技有限公司 SHANGHAI CAMBRICON INFORMATION TECHNOLOGY CO., LTD. [CN]/[CN]
Inventors
  • 陈天石 CHEN, Tianshi
  • 刘少礼 LIU, Shaoli
  • 王在 WANG, Zai
  • 胡帅 HU, Shuai
Agents
  • 广州三环专利商标代理有限公司 SCIHEAD IP LAW FIRM
Priority Data
201710595951.120.07.2017CN
Publication Language Chinese (ZH)
Filing Language Chinese (ZH)
Designated States
Title
(EN) CALCULATION METHOD AND RELATED PRODUCT
(FR) PROCÉDÉ DE CALCUL ET PRODUIT ASSOCIÉ
(ZH) 一种计算方法及相关产品
Abstract
(EN)
Provided by the present disclosure is a calculation method, said method being applied in a calculation apparatus, wherein said calculation apparatus comprises: a memory, a register unit, and a matrix calculation unit; and said method comprises the following steps: said calculation apparatus controlling said matrix calculation unit to acquire a first operation command, said first operation command comprising a matrix reading instruction required to execute said command; said calculation apparatus controlling an operation unit to send a reading command to the memory according to said matrix reading instruction; said calculation apparatus controlling said operation unit to read a matrix corresponding to said matrix reading instruction on the basis of employing a batch reading mode, and executing said first operation command for the matrix. The technical solution provided by the present application has the advantages of performing calculations quickly and being highly efficient.
(FR)
L'invention concerne un procédé de calcul, ledit procédé étant appliqué dans un appareil de calcul, ledit appareil de calcul comprenant une mémoire, une unité de registre et une unité de calcul de matrice. Ledit procédé comprend les étapes suivantes : ledit appareil de calcul amène ladite unité de calcul de matrice à acquérir une première commande de fonctionnement, ladite première commande de fonctionnement comprenant une instruction de lecture de matrice requise pour exécuter ladite commande; ledit appareil de calcul amène une unité de fonctionnement à envoyer une commande de lecture à la mémoire en fonction de ladite instruction de lecture de matrice; ledit appareil de calcul amène ladite unité de fonctionnement à lire une matrice correspondant à ladite instruction de lecture de matrice d'après l'utilisation d'un mode de lecture par lots, puis exécute ladite première commande de fonctionnement pour la matrice. La solution technique de l'invention offre l'avantage de pouvoir effectuer des calculs rapidement et d'être très efficace.
(ZH)
本公开提供了一种计算方法,所述方法应用于计算装置内,所述计算装置包括:存储器、寄存器单元和矩阵计算单元;所述方法包括如下步骤:所述计算装置控制所述矩阵计算单元获取第一运算指令,所述第一运算指令包括执行所述指令所需的矩阵读取指示;所述计算装置控制所述运算单元依据所述矩阵读取指示向所述存储器发送读取命令;所述计算装置控制所述运算单元依据采用批量读取方式读取所述矩阵读取指示对应的矩阵,对该矩阵执行所述第一运算指令。本申请提供的技术方案具有计算速度快,效率高的优点。
Latest bibliographic data on file with the International Bureau