WO2021120711A8 - 一种矩阵乘法器、数据处理方法、集成电路器件及处理器 - Google Patents

一种矩阵乘法器、数据处理方法、集成电路器件及处理器 Download PDF

Info

Publication number
WO2021120711A8
WO2021120711A8 PCT/CN2020/114000 CN2020114000W WO2021120711A8 WO 2021120711 A8 WO2021120711 A8 WO 2021120711A8 CN 2020114000 W CN2020114000 W CN 2020114000W WO 2021120711 A8 WO2021120711 A8 WO 2021120711A8
Authority
WO
WIPO (PCT)
Prior art keywords
matrix
elements
vgprs
vsps
processor
Prior art date
Application number
PCT/CN2020/114000
Other languages
English (en)
French (fr)
Other versions
WO2021120711A1 (zh
Inventor
左航
Original Assignee
成都海光微电子技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 成都海光微电子技术有限公司 filed Critical 成都海光微电子技术有限公司
Publication of WO2021120711A1 publication Critical patent/WO2021120711A1/zh
Publication of WO2021120711A8 publication Critical patent/WO2021120711A8/zh

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F7/00—Methods or arrangements for processing data by operating upon the order or content of the data handled
    • G06F7/38—Methods or arrangements for performing computations using exclusively denominational number representation, e.g. using binary, ternary, decimal representation
    • G06F7/48—Methods or arrangements for performing computations using exclusively denominational number representation, e.g. using binary, ternary, decimal representation using non-contact-making devices, e.g. tube, solid state device; using unspecified devices
    • G06F7/52—Multiplying; Dividing
    • G06F7/523—Multiplying only

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Mathematical Physics (AREA)
  • General Engineering & Computer Science (AREA)
  • Computing Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Algebra (AREA)
  • Databases & Information Systems (AREA)
  • Software Systems (AREA)
  • Complex Calculations (AREA)
  • Advance Control (AREA)

Abstract

一种矩阵乘法器、数据处理方法、集成电路器件及处理器。该矩阵乘法器包括：LDS，被配置成按照行顺序存储第一矩阵；K个VGPR，被配置成存储第二矩阵中的各个列，每个VGPR存储第二矩阵的一列；与K个VGPR一一对应连接的K个VSP，LDS通过总线与每个VSP均连接，使得第一矩阵中的元素逐个并行地被加载到K个VSP，与K个VGPR中各自存储的列对应的元素进行相乘，K个VSP并行地将第一矩阵同一行中的元素逐个与第二矩阵的对应元素产生的相乘结果依次累加，得到第三矩阵同一行的所有元素，从而完成第一矩阵和第二矩阵的乘法运算。该矩阵乘法器可以针对第三矩阵同一行的所有元素进行并行计算，从而减少从第一矩阵中获取元素的次数。
PCT/CN2020/114000 2019-12-16 2020-09-08 一种矩阵乘法器、数据处理方法、集成电路器件及处理器 WO2021120711A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201911302512.2A CN111079081B (zh) 2019-12-16 2019-12-16 一种矩阵乘法器、数据处理方法、集成电路器件及处理器
CN201911302512.2 2019-12-16

Publications (2)

Publication Number Publication Date
WO2021120711A1 WO2021120711A1 (zh) 2021-06-24
WO2021120711A8 true WO2021120711A8 (zh) 2021-08-05

Family

ID=70315128

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/114000 WO2021120711A1 (zh) 2019-12-16 2020-09-08 一种矩阵乘法器、数据处理方法、集成电路器件及处理器

Country Status (2)

Country Link
CN (1) CN111079081B (zh)
WO (1) WO2021120711A1 (zh)

Families Citing this family (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111079081B (zh) * 2019-12-16 2021-02-12 海光信息技术股份有限公司 一种矩阵乘法器、数据处理方法、集成电路器件及处理器
CN112182496B (zh) * 2020-09-24 2022-09-16 成都海光集成电路设计有限公司 用于矩阵乘法的数据处理方法及装置
CN112506567B (zh) * 2020-11-27 2022-11-04 海光信息技术股份有限公司 数据读取方法和数据读取电路
CN112433760B (zh) * 2020-11-27 2022-09-23 海光信息技术股份有限公司 数据排序方法和数据排序电路
CN112434256B (zh) * 2020-12-03 2022-09-13 海光信息技术股份有限公司 矩阵乘法器和处理器
CN115880132B (zh) * 2023-02-06 2023-05-23 南京砺算科技有限公司 图形处理器、矩阵乘法任务处理方法、装置及存储介质
CN116109468B (zh) * 2023-04-04 2023-07-21 南京砺算科技有限公司 图形处理单元及指令编译方法、存储介质、终端设备

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5784636A (en) * 1996-05-28 1998-07-21 National Semiconductor Corporation Reconfigurable computer architecture for use in signal processing applications
CN102375721B (zh) * 2010-08-23 2016-03-30 联想(北京)有限公司 一种矩阵乘法运算方法、图形处理器和电子设备
CN102510329B (zh) * 2011-09-29 2014-08-13 中国人民解放军信息工程大学 一种乘法器及其控制方法
CN102662623A (zh) * 2012-04-28 2012-09-12 电子科技大学 基于单fpga的并行矩阵乘法器及其实现方法
US9384168B2 (en) * 2013-06-11 2016-07-05 Analog Devices Global Vector matrix product accelerator for microprocessor integration
CN109992743B (zh) * 2017-12-29 2020-06-16 华为技术有限公司 矩阵乘法器
CN111079081B (zh) * 2019-12-16 2021-02-12 海光信息技术股份有限公司 一种矩阵乘法器、数据处理方法、集成电路器件及处理器

Also Published As

Publication number Publication date
CN111079081B (zh) 2021-02-12
WO2021120711A1 (zh) 2021-06-24
CN111079081A (zh) 2020-04-28

Similar Documents

Publication Publication Date Title
WO2021120711A8 (zh) 一种矩阵乘法器、数据处理方法、集成电路器件及处理器
US11816532B2 (en) Performing kernel striding in hardware
US11361051B1 (en) Dynamic partitioning
US10635965B2 (en) Neural network convolution computation method and device, and computer-readable storage medium
US10915297B1 (en) Hardware accelerator for systolic matrix multiplication
US8051124B2 (en) High speed and efficient matrix multiplication hardware module
CN106844294B (zh) 卷积运算芯片和通信设备
US20190095776A1 (en) Efficient data distribution for parallel processing
US10713214B1 (en) Hardware accelerator for outer-product matrix multiplication
EP3803638A1 (en) Processing matrix vector multiplication
GB2582094A (en) Matrix computation engine
CN109144469B (zh) 流水线结构神经网络矩阵运算架构及方法
US20220350861A1 (en) Systolic array and accelerator including the same
WO2003043236A1 (en) Array processing for linear system solutions
GB2601701A (en) Performing dot product operations using a memristive crossbar array
EP4095719A1 (en) Sparse matrix multiplication in hardware
US6598061B1 (en) System and method for performing modular multiplication
US20080228845A1 (en) Apparatus for calculating an n-point discrete fourier transform by utilizing cooley-tukey algorithm
CN113743046B (zh) 存算一体版图结构和数据拆分存算一体版图结构
JP5269137B2 (ja) 演算装置
US20150154005A1 (en) Methods and Apparatuses for Performing Multiplication
US20240281209A1 (en) Machine learning optimization circuit and method thereof
CN110647976B (zh) 一种矩阵卷积优化运算方法和电路
US20240220203A1 (en) Streaming-based compute unit and method, and artificial intelligence chip
US20230176819A1 (en) Pipelined processing of polynomial computation

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20901515

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20901515

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 20901515

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 270323)

122 Ep: pct application non-entry in european phase

Ref document number: 20901515

Country of ref document: EP

Kind code of ref document: A1