CN112703511A - 运算加速器和数据处理方法 - Google Patents

运算加速器和数据处理方法 Download PDF

Info

Publication number
CN112703511A
CN112703511A CN201880097652.8A CN201880097652A CN112703511A CN 112703511 A CN112703511 A CN 112703511A CN 201880097652 A CN201880097652 A CN 201880097652A CN 112703511 A CN112703511 A CN 112703511A
Authority
CN
China
Prior art keywords
matrix
memory
convolution
converter
row
Prior art date
Legal status (The legal status 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 status listed.)
Granted
Application number
CN201880097652.8A
Other languages
English (en)
Other versions
CN112703511B (zh
Inventor
顾雄礼
李艳华
张惠敏
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huawei Technologies Co Ltd
Original Assignee
Huawei Technologies Co Ltd
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 Huawei Technologies Co Ltd filed Critical Huawei Technologies Co Ltd
Publication of CN112703511A publication Critical patent/CN112703511A/zh
Application granted granted Critical
Publication of CN112703511B publication Critical patent/CN112703511B/zh
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/06—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
    • G06N3/063—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons using electronic means
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • General Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Computational Linguistics (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Artificial Intelligence (AREA)
  • Neurology (AREA)
  • Complex Calculations (AREA)

Abstract

一种运算加速器和数据处理方法，该运算加速器包括：第一存储器，用于存储第一矩阵，所述第一矩阵的每一行或每一列为第二矩阵的一个子矩阵转换的向量，所述第二矩阵为损失函数的输出值对卷积层输出的卷积结果的偏导数；第二存储器，用于存储第三矩阵，所述第三矩阵为第四矩阵经过转置和反向排列得到的矩阵，所述第四矩阵为在所述卷积层执行卷积运算得到所述卷积结果所使用的权重矩阵；分别与所述第一存储器、所述第二存储器连接的运算电路；所述运算电路，用于获取所述第一矩阵和所述第三矩阵，计算所述第一矩阵和所述第三矩阵的乘积，得到第五矩阵；不需要col2img操作，就可以计算出损失函数对输入矩阵的偏导数，计算效率高。

Description

PCT国内申请，说明书已公开。

Claims (15)

  1. PCT国内申请，权利要求书已公开。
CN201880097652.8A 2018-09-27 2018-09-27 运算加速器和数据处理方法 Active CN112703511B (zh)

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2018/108018 WO2020061924A1 (zh) 2018-09-27 2018-09-27 运算加速器和数据处理方法

Publications (2)

Publication Number Publication Date
CN112703511A true CN112703511A (zh) 2021-04-23
CN112703511B CN112703511B (zh) 2023-08-25

Family

ID=69950853

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201880097652.8A Active CN112703511B (zh) 2018-09-27 2018-09-27 运算加速器和数据处理方法

Country Status (2)

Country Link
CN (1) CN112703511B (zh)
WO (1) WO2020061924A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114936633A (zh) * 2022-06-15 2022-08-23 北京爱芯科技有限公司 用于转置运算的数据处理单元及图像转置运算方法
CN116861149A (zh) * 2023-09-05 2023-10-10 之江实验室 卷积运算的优化方法、装置及处理器

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113610221B (zh) * 2021-06-29 2024-02-13 西安电子科技大学 一种基于fpga的可变膨胀卷积运算硬件系统
CN116108902B (zh) * 2023-02-22 2024-01-05 成都登临科技有限公司 采样操作实现系统、方法、电子设备及存储介质
CN117291240B (zh) * 2023-11-24 2024-03-15 芯来智融半导体科技(上海)有限公司 卷积神经网络加速器及电子设备

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104915322A (zh) * 2015-06-09 2015-09-16 中国人民解放军国防科学技术大学 一种卷积神经网络硬件加速方法及其axi总线ip核
US20170200078A1 (en) * 2014-08-28 2017-07-13 Commissariat A L'energie Atomique Et Aux Energies Alternatives Convolutional neural network
CN107665365A (zh) * 2016-07-27 2018-02-06 三星电子株式会社 卷积神经网络中的加速器及其操作方法
CN108205687A (zh) * 2018-02-01 2018-06-26 通号通信信息集团有限公司 目标检测系统中基于关注点机制定位损失计算方法及系统
CN108241484A (zh) * 2016-12-26 2018-07-03 上海寒武纪信息科技有限公司 基于高带宽存储器的神经网络计算装置和方法
CN108320026A (zh) * 2017-05-16 2018-07-24 腾讯科技（深圳）有限公司 机器学习模型训练方法和装置

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105892989B (zh) * 2016-03-28 2017-04-12 中国科学院计算技术研究所 一种神经网络加速器及其运算方法
CN108416433B (zh) * 2018-01-22 2020-11-24 上海熠知电子科技有限公司 一种基于异步事件的神经网络异构加速方法和系统

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170200078A1 (en) * 2014-08-28 2017-07-13 Commissariat A L'energie Atomique Et Aux Energies Alternatives Convolutional neural network
CN104915322A (zh) * 2015-06-09 2015-09-16 中国人民解放军国防科学技术大学 一种卷积神经网络硬件加速方法及其axi总线ip核
CN107665365A (zh) * 2016-07-27 2018-02-06 三星电子株式会社 卷积神经网络中的加速器及其操作方法
CN108241484A (zh) * 2016-12-26 2018-07-03 上海寒武纪信息科技有限公司 基于高带宽存储器的神经网络计算装置和方法
CN108320026A (zh) * 2017-05-16 2018-07-24 腾讯科技（深圳）有限公司 机器学习模型训练方法和装置
CN108205687A (zh) * 2018-02-01 2018-06-26 通号通信信息集团有限公司 目标检测系统中基于关注点机制定位损失计算方法及系统

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114936633A (zh) * 2022-06-15 2022-08-23 北京爱芯科技有限公司 用于转置运算的数据处理单元及图像转置运算方法
CN116861149A (zh) * 2023-09-05 2023-10-10 之江实验室 卷积运算的优化方法、装置及处理器
CN116861149B (zh) * 2023-09-05 2024-01-09 之江实验室 卷积运算的优化方法、装置及处理器

Also Published As

Publication number Publication date
WO2020061924A1 (zh) 2020-04-02
CN112703511B (zh) 2023-08-25

Similar Documents

Publication Publication Date Title
CN112703511B (zh) 运算加速器和数据处理方法
US20210224125A1 (en) Operation Accelerator, Processing Method, and Related Device
EP3373210B1 (en) Transposing neural network matrices in hardware
EP3637281A1 (en) Operational accelerator
US20190212982A1 (en) Processor, information processing apparatus and operation method for processor
JP2019125352A (ja) 畳み込みニューラルネットワークの畳み込み層における演算を実行する方法及び装置
JP6713036B2 (ja) 折り畳まれた特徴データに対して畳み込み演算を実行するための方法および装置
CN111461311B (zh) 基于众核处理器的卷积神经网络运算加速方法及装置
CN108573305B (zh) 一种数据处理方法、设备及装置
CN108629406B (zh) 用于卷积神经网络的运算装置
CN113222101A (zh) 深度学习处理装置、方法、设备和存储介质
CN108170640B (zh) 神经网络运算装置及应用其进行运算的方法
WO2022067508A1 (zh) 一种神经网络加速器、加速方法以及装置
US11120101B2 (en) Matrix multiplication system and method
JP2020126597A (ja) 計算装置と計算方法
CN114995782B (zh) 数据处理方法、装置、设备和可读存储介质
CN109993293B (zh) 一种适用于堆叠式沙漏网络的深度学习加速器
CN110580519A (zh) 一种卷积运算结构及其方法
CN112559043A (zh) 一种轻量级人工智能加速模块
CN112789627B (zh) 一种神经网络处理器、数据处理方法及相关设备
WO2022151779A1 (zh) 卷积运算的实现方法、数据处理方法及装置
WO2022007265A1 (zh) 一种膨胀卷积加速计算方法及装置
WO2022000225A1 (zh) 一种卷积神经网络数据处理方法及其相关设备
WO2023122896A1 (zh) 一种数据处理方法和装置
US11526305B2 (en) Memory for an artificial neural network accelerator

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant