FR3117626B1 - Procédé d’allocation de mémoire pendant l’exécution d’un réseau de neurones - Google Patents

Procédé d’allocation de mémoire pendant l’exécution d’un réseau de neurones Download PDF

Info

Publication number
FR3117626B1
FR3117626B1 FR2013373A FR2013373A FR3117626B1 FR 3117626 B1 FR3117626 B1 FR 3117626B1 FR 2013373 A FR2013373 A FR 2013373A FR 2013373 A FR2013373 A FR 2013373A FR 3117626 B1 FR3117626 B1 FR 3117626B1
Authority
FR
France
Prior art keywords
execution
neural network
layers
placements
order
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.)
Active
Application number
FR2013373A
Other languages
English (en)
Other versions
FR3117626A1 (fr
Inventor
Laurent Folliot
Mirko Falchetto
Pierre Demaj
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.)
STMicroelectronics Rousset SAS
STMicroelectronics SRL
Original Assignee
STMicroelectronics Rousset SAS
STMicroelectronics SRL
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 STMicroelectronics Rousset SAS, STMicroelectronics SRL filed Critical STMicroelectronics Rousset SAS
Priority to FR2013373A priority Critical patent/FR3117626B1/fr
Priority to US17/455,770 priority patent/US20220188610A1/en
Priority to EP21211915.0A priority patent/EP4016311B1/fr
Priority to CN202111534114.0A priority patent/CN114638350A/zh
Publication of FR3117626A1 publication Critical patent/FR3117626A1/fr
Application granted granted Critical
Publication of FR3117626B1 publication Critical patent/FR3117626B1/fr
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F12/00—Accessing, addressing or allocating within memory systems or architectures
    • G06F12/02—Addressing or allocation; Relocation
    • G06F12/0223—User address space allocation, e.g. contiguous or non contiguous base addressing
    • G06F12/023—Free address space management
    • 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
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
    • G06F3/0601—Interfaces specially adapted for storage systems
    • G06F3/0602—Interfaces specially adapted for storage systems specifically adapted to achieve a particular effect
    • G06F3/0604—Improving or facilitating administration, e.g. storage management
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
    • G06F3/0601—Interfaces specially adapted for storage systems
    • G06F3/0628—Interfaces specially adapted for storage systems making use of a particular technique
    • G06F3/0638—Organizing or formatting or addressing of data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
    • G06F3/0601—Interfaces specially adapted for storage systems
    • G06F3/0668—Interfaces specially adapted for storage systems adopting a particular infrastructure
    • G06F3/0671—In-line storage system
    • 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
    • G06N3/04—Architecture, e.g. interconnection topology
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2212/00—Indexing scheme relating to accessing, addressing or allocation within memory systems or architectures
    • G06F2212/10—Providing a specific technical effect
    • G06F2212/1041—Resource optimization
    • G06F2212/1044—Space efficiency improvement
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2212/00—Indexing scheme relating to accessing, addressing or allocation within memory systems or architectures
    • G06F2212/20—Employing a main memory using a specific memory technology

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Biophysics (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Artificial Intelligence (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Computational Linguistics (AREA)
  • Evolutionary Computation (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Neurology (AREA)
  • Human Computer Interaction (AREA)
  • Memory System Of A Hierarchy Structure (AREA)
  • Memory System (AREA)
  • Devices For Executing Special Programs (AREA)

Abstract

Selon un aspect, un procédé est proposé pour définir des placements, dans une mémoire volatile, de mémoires tampons de travail temporaires utilisées pendant une exécution d’un réseau de neurones artificiels, le procédé comprenant les étapes suivantes : - déterminer (20) un ordre d’exécution des couches du réseau de neurones, - définir (21) des placements, dans une zone de mémoire de tas de la mémoire volatile, de mémoires tampons de résultats intermédiaires générées par chaque couche, selon l’ordre d’exécution des couches, - déterminer (22) au moins une région libre de la zone de mémoire de tas pendant l’exécution des couches, - définir (27) des placements de mémoires tampons de travail temporaires dans ladite au moins une région libre de la zone de mémoire de tas selon l’ordre d’exécution des couches. Figure pour l’abrégé : Fig 2
FR2013373A 2020-12-16 2020-12-16 Procédé d’allocation de mémoire pendant l’exécution d’un réseau de neurones Active FR3117626B1 (fr)

Priority Applications (4)

Application Number Priority Date Filing Date Title
FR2013373A FR3117626B1 (fr) 2020-12-16 2020-12-16 Procédé d’allocation de mémoire pendant l’exécution d’un réseau de neurones
US17/455,770 US20220188610A1 (en) 2020-12-16 2021-11-19 Method for memory allocation during execution of a neural network
EP21211915.0A EP4016311B1 (fr) 2020-12-16 2021-12-02 Procédé d'attribution de mémoire pendant l'exécution d'un réseau neuronal
CN202111534114.0A CN114638350A (zh) 2020-12-16 2021-12-15 用于神经网络执行期间的存储器分配的方法

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
FR2013373A FR3117626B1 (fr) 2020-12-16 2020-12-16 Procédé d’allocation de mémoire pendant l’exécution d’un réseau de neurones
FR2013373 2020-12-16

Publications (2)

Publication Number Publication Date
FR3117626A1 FR3117626A1 (fr) 2022-06-17
FR3117626B1 true FR3117626B1 (fr) 2023-12-01

Family

ID=75438898

Family Applications (1)

Application Number Title Priority Date Filing Date
FR2013373A Active FR3117626B1 (fr) 2020-12-16 2020-12-16 Procédé d’allocation de mémoire pendant l’exécution d’un réseau de neurones

Country Status (4)

Country Link
US (1) US20220188610A1 (fr)
EP (1) EP4016311B1 (fr)
CN (1) CN114638350A (fr)
FR (1) FR3117626B1 (fr)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20240020897A1 (en) * 2022-07-12 2024-01-18 Nvidia Corporation Neural network-based image lighting

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE1768002A1 (de) 1968-03-20 1972-03-16 Hoechst Ag Harnstoffderivate,ihre Herstellung sowie Anwendung als Herbicide
US10699186B2 (en) * 2015-12-02 2020-06-30 Google Llc Determining orders of execution of a neural network
KR102561261B1 (ko) * 2017-11-14 2023-07-28 삼성전자주식회사 커널을 이용한 컨볼루션 연산 처리 방법 및 장치
FR3089649A1 (fr) 2018-12-06 2020-06-12 Stmicroelectronics (Rousset) Sas Procédé et dispositif de détermination de la taille mémoire globale d’une zone mémoire globale allouée aux données d’un réseau de neurones
FR3094104A1 (fr) * 2019-03-20 2020-09-25 Stmicroelectronics (Rousset) Sas Procédé et dispositif de détermination de la taille mémoire globale d’une zone mémoire globale allouée aux données d’un réseau de neurones compte tenu de sa topologie
CN110162338B (zh) * 2019-05-31 2020-08-04 中科寒武纪科技股份有限公司 运算方法、装置及相关产品

Also Published As

Publication number Publication date
FR3117626A1 (fr) 2022-06-17
CN114638350A (zh) 2022-06-17
EP4016311B1 (fr) 2024-05-01
EP4016311A1 (fr) 2022-06-22
US20220188610A1 (en) 2022-06-16

Similar Documents

Publication Publication Date Title
FR3117626B1 (fr) Procédé d’allocation de mémoire pendant l’exécution d’un réseau de neurones
Johannesson et al. The cost‐effectiveness of treating hypertension in elderly people—an analysis of the Swedish Trial in Old Patients with Hypertension (STOP Hypertension)
Newcombe et al. Sympathectomy for hypertension
Onwochei et al. Carbetocin reduces the need for additional uterotonics in elective caesarean delivery: a systematic review, meta-analysis and trial sequential analysis of randomised controlled trials
Pastori et al. Temporal trends of time in therapeutic range and incidence of cardiovascular events in patients with non-valvular atrial fibrillation
Lindsay et al. Failure of response of menopausal vasomotor symptoms to clonidine
Nigora et al. System of Statistic Indicators of Regional Sustainable Development in Conditions of Digital Econemy
Botsile et al. Incidence and risk factors for thromboembolism and major bleeding in patients with mechanical heart valves: a tertiary hospital-based study in Botswana
Bühler et al. β-Blockers and calcium antagonists: cornerstones of antihypertensive therapy in the 1980s
Peres et al. Does Lip‐Repositioning Surgery Improve Long‐Term Smile Outcome and Dental Esthetics in Patients With Excessive Gingival Display? A Review of the Current Literature
EP3657534A3 (fr) Structure de boîtier de semi-conducteur et son procédé de fabrication
Bulliard et al. Breast cancer screening programmes in Switzerland, 2010–2018
FR3109833B1 (fr) Procédé d’allocation de données intermédiaires d’un réseau de neurones artificiel
Zuniga et al. Does Immediate Long-Span Nerve Allograft Reconstruction Affect the Incidence of Chronic Postsurgical and Neuropathic Pain in the Reconstructed Mandible Following Resection for Benign and Malignant Disease
Law et al. A pharmacoeconomic study of traditional anticoagulation versus direct oral anticoagulation for the treatment of venous thromboembolism in the emergency department
Fendrick A pharmacoeconomic perspective on stroke prevention in atrial fibrillation
Okada et al. Is High Heart Rate Always Harmful to Heart Failure Patients?
Baghel et al. Comparison of open hemorrhoidectomy under local and spinal anesthesia and its practical feasibility at a tertiary care institute
Caraher Reproductive Injustice? A County-Level Analysis of the Impact of Abortion Restrictions on Abortion Rates
Tilman Atrial fibrillation: In the light of new hypothesis
Cheah et al. Relationship between clinicopathological characteristics and margin involvement after SSM versus SIM
Pastore Implementing nurse led telephone clinics in support of the personalised care pathway
Zovi et al. 2SPD-024 Shortages of medicines in hospital: results of a survey on the perception of health workers in the wards VSV real world
Jaakkola et al. Subtype of atrial fibrillation and the outcome of transcatheter aortic valve replacement: The FinnValve Study
Alam et al. Prescribing Pattern of Antihypertensive Drugs in Hypertensive Patients in Tertiary Care Hospital

Legal Events

Date Code Title Description
PLFP Fee payment

Year of fee payment: 2

PLSC Publication of the preliminary search report

Effective date: 20220617

PLFP Fee payment

Year of fee payment: 3

PLFP Fee payment

Year of fee payment: 4