Cover Image
市場調查報告書

全球深度學習晶片組市場分析與預測:CPU、GPU、FPGA、ASIC、其他晶片組

Deep Learning Chipsets - CPUs, GPUs, FPGAs, ASICs and Other Chipsets for Training, and Inference Applications: Global Market Analysis and Forecasts

出版商 Tractica 商品編碼 408667
出版日期 內容資訊 英文 81 Pages; 139 Tables, Charts & Figures
訂單完成後即時交付
價格
Back to Top
全球深度學習晶片組市場分析與預測:CPU、GPU、FPGA、ASIC、其他晶片組 Deep Learning Chipsets - CPUs, GPUs, FPGAs, ASICs and Other Chipsets for Training, and Inference Applications: Global Market Analysis and Forecasts
出版日期: 2017年03月03日 內容資訊: 英文 81 Pages; 139 Tables, Charts & Figures
簡介

深度學習晶片組市場的收益預計從2016年的5億1300萬美元,到2025年增加到122億美元,2016年∼2025年以42.2%的年複合成長率成長。

本報告提供全球深度學習晶片組市場相關調查分析,市場課題,技術課題,主要企業,市場預測等系統性資訊。

第1章 摘要整理

第2章 市場課題

  • 推動市場要素
  • 市場障礙與課題
  • 主要市場與用途
  • 地區差

第3章 技術課題

  • 人工智能和深度學習
  • 深度學習:主要概念和晶片組的影響
  • 數理問題的深度學習
  • 訓練 vs. 推論
  • 有監視 vs. 無監視的訓練
  • 深度學習晶片組
  • 低準確度,固定小數點,浮點
  • 深度學習開發的組成架構
  • OpenCL和CUDA
  • 深度學習工作站
  • 神經形態處理
  • 大學、研究機關

第4章 主要參與企業

  • Google
  • Intel
  • Tera深度
  • Xilinx
  • AMD
  • NVIDIA
  • ARM
  • Qualcomm
  • IBM
  • Graphcore
  • BrainChip
  • Knowm
  • KnuEdge
  • Mobileye
  • Artificial Learning
  • Wave Computing
  • CEVA
  • Movidius
  • Nervana Systems
  • Amazon
  • Facebook

第5章 市場預測

  • 預測手法與前提條件
  • 市場全體
  • 出貨數:晶片組的各類型
  • 收益:晶片組的各類型
  • 收益:訓練 vs. 推論
  • 收益:各電腦容量
  • 收益:各電力消耗
  • 平均銷售價格:晶片組的各類型
  • 收益:各市場部門
  • CPU
  • GPU
  • ASIC
  • FPGA
  • 其他晶片組
  • 結論

第6章 企業名錄

第7章 簡稱、首字母

第8章 目錄

第9章 圖表

第10章 調查範圍、資訊來源、調查手法、註記

目錄
Product Code: DLC-17

Deep learning technology is driving the evolution of artificial intelligence (AI) and has become one of the hottest topics of discussion within the technology world and beyond. Given the rate at which deep learning is progressing, some industry observers are predicting it will bring about a doomsday scenario, while others strive for a time when the technology can transform business processes and create new business models through scalable, efficient automation and predictive capabilities. Against the backdrop of rapid technological development, the current market climate is ripe for innovation in hardware in general, and chipsets more specifically. The chipset market today is led by graphics processing units (GPUs) and central processing units (CPUs), but during the coming years there will be an expanded role for other chipset types including application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), and other emerging chipsets.

Deep learning technology has primarily been a software play so far. The need for hardware acceleration was recognized in academia, but few companies ventured into developing specialized chipsets until now. AI applications demand higher performance and lower power. The world's top semiconductor companies as well as a number of startups have jumped into the race to meet these demands and this report attempts to capture state-of-the-art offerings currently available in the industry. Tractica forecasts that the deep learning chipset market revenue will increase from $513 million in 2016 to $12.2 billion by 2025 at a compound annual growth rate (CAGR) of 42.2%.

This Tractica report discusses deep learning chipsets from the points of view of both the market and the technology. Tractica segments the application markets, as well as the chipsets themselves by category, and provides market forecasts for the period from 2016 through 2025. Deep learning technology as it pertains to chipsets is very technical and Tractica explains the various approaches different vendors are taking to solve the problem and the tradeoffs involved. A comprehensive listing of the companies in this market is also provided.

Key Market Forecasts

  • Deep Learning Chipset Unit Shipments and Revenue by Type, World Markets: 2016-2025
  • Deep Learning Chipset Revenue, Inference versus Training, World Markets: 2016-2025
  • Deep Learning Chipset Revenue by Compute Capacity, World Markets: 2016-2025
  • Deep Learning Chipset Revenue by Power Consumption, World Markets: 2016-2025
  • Deep Learning Chipset Average Selling Price by Type, World Markets: 2016-2025
  • Deep Learning Chipset Revenue by Market Sector, World Markets: 2016-2025
  • Deep Learning CPU Unit Shipments and Revenue, World Markets: 2016-2025
  • Deep Learning GPU Unit Shipments and Revenue, World Markets: 2016-2025
  • Deep Learning ASIC Unit Shipments and Revenue, World Markets: 2016-2025
  • Deep Learning FPGA Unit Shipments and Revenue, World Markets: 2016-2025
  • Deep Learning Other Chipset Unit Shipments and Revenue, World Markets: 2016-2025

Chipset Types

  • CPU
  • GPU
  • FPGA
  • ASIC
  • Other

Power Consumption

  • High (>100 W)
  • Medium (5-100 W)
  • Low (<5 W)

Compute Capacity

  • High (> 1TFlops)
  • Low (< 1TFlops)

Application Domains

  • Training
  • Inference
  • Inference & Training

Market Sectors

  • Enterprise
  • Consumer

Geographies

  • North America
  • Europe
  • Asia Pacific
  • Latin America
  • Middle East and Africa

image1

Key Questions Addressed

  • What is the mix of chipset types being used for deep learning today, and how will it change during the next 10 years?
  • Which chipset types are most appropriate for training versus inference applications?
  • What will be the power consumption and compute capacity profiles of chipsets used for various deep learning applications?
  • What is the market opportunity for deep learning chipsets in cloud/data center environments versus edge devices?
  • Which market sectors and industries will drive demand for deep learning chipsets?
  • What is the state of technology development for deep learning chipsets, and who are the key industry players driving innovation?

Who Needs This Report?

  • Semiconductor and component manufacturers
  • Service providers and systems integrators
  • End-user organizations deploying deep learning systems
  • Industry associations
  • Government agencies
  • Investor community

Table of Contents

1. Executive Summary

  • 1.1. Market Development
  • 1.2. Background
  • 1.3. Critique of Deep Learning in Hardware
  • 1.4. Market Segmentation
    • 1.4.1. Segmentation by Architecture
    • 1.4.2. Segmentation Based on Training versus Inference
    • 1.4.3. Segmentation Based on Compute Capacity
    • 1.4.4. Segmentation Based on Power Consumption
    • 1.4.5. Neuromorphic Chipsets
  • 1.5. Technology Parameters
    • 1.5.1. Performance
    • 1.5.2. Power
    • 1.5.3. Performance per Watt
    • 1.5.4. Programmability
    • 1.5.5. Intellectual Property and Ecosystem
    • 1.5.6. Development Tools
    • 1.5.7. Floating and Fixed Point Support
  • 1.6. Market Forecasts

2. Market Issues

  • 2.1. Market Drivers
    • 2.1.1. Exponential Growth in the Number of Enterprise Artificial Intelligence Applications
    • 2.1.2. Computer Vision Applications
    • 2.1.3. Embedded Devices
    • 2.1.4. Accelerated Computing
  • 2.2. Market Barriers and Challenges
    • 2.2.1. Development Costs
    • 2.2.2. Availability of Expertise
    • 2.2.3. Time to Market
  • 2.3. Key Markets and Applications
    • 2.3.1. Consumer
    • 2.3.2. Enterprise
    • 2.3.3. Industrial
    • 2.3.4. Transportation
    • 2.3.5. Defense
    • 2.3.6. Other
  • 2.4. Regional Differences

3. Technology Issues

  • 3.1. Artificial Intelligence and Deep Learning
  • 3.2. Deep Learning: Key Concepts and Implications for Chipsets
    • 3.2.1. Feedforward Networks versus Recurrent Networks
    • 3.2.2. Convolutional Neural Network
    • 3.2.3. Long Short-Term Memory
  • 3.3. Deep Learning in Terms of Mathematical Problems
    • 3.3.1. Classification
    • 3.3.2. Regression
    • 3.3.3. Transcription
    • 3.3.4. Machine Translation
    • 3.3.5. Anomaly Detection
  • 3.4. Training versus Inference
  • 3.5. Supervised versus Unsupervised Training
  • 3.6. Chipsets for Deep Learning
    • 3.6.1. Central Processing Units
    • 3.6.2. Graphics Processing Units
    • 3.6.3. Field Programmable Gate Arrays
    • 3.6.4. Application Specific Integrated Circuits
  • 3.7. Low Precision, Fixed Point, and Floating Point
  • 3.8. Deep Learning Development Frameworks
  • 3.9. OpenCL and CUDA
  • 3.10. Deep Learning Workstations
  • 3.11. Neuromorphic Processing
  • 3.12. Universities and Research Institutions

4. Key Industry Players

  • 4.1. Google
  • 4.2. Intel
  • 4.3. TeraDeep
  • 4.4. Xilinx
  • 4.5. AMD
  • 4.6. NVIDIA
  • 4.7. ARM
  • 4.8. Qualcomm
  • 4.9. IBM
  • 4.10. Graphcore
  • 4.11. BrainChip
  • 4.12. Knowm
  • 4.13. KnuEdge
  • 4.14. Mobileye
  • 4.15. Artificial Learning
  • 4.16. Wave Computing
  • 4.17. CEVA
  • 4.18. Movidius
  • 4.19. Nervana Systems
  • 4.20. Amazon
  • 4.21. Facebook

5. Market Forecasts

  • 5.1. Forecast Methodology and Assumptions
  • 5.2. Overall Market
  • 5.3. Unit Shipments by Chipset Type
  • 5.4. Revenue by Chipset Type
  • 5.5. Revenue by Training versus Inference
  • 5.6. Revenue by Compute Capacity
  • 5.7. Revenue by Power Consumption
  • 5.8. Average Selling Price by Chipset Type
  • 5.9. Revenue by Market Sector
  • 5.10. Central Processing Units
  • 5.11. Graphics Processing Units
  • 5.12. Application-Specific Integrated Circuits
  • 5.13. Field Programmable Gate Arrays
  • 5.14. Other Chipsets
  • 5.15. Conclusion

6. Company Directory

7. Acronym and Abbreviation List

8. Table of Contents

9. Table of Charts and Figures

10. Scope of Study, Sources and Methodology, Notes

Charts

  • Deep Learning Chipset Revenue by Type, World Markets: 2016-2025
  • Deep Learning Chipset Revenue by Region, World Markets: 2016-2025
  • Deep Learning Chipset Year-on-Year Revenue Growth Rates, World Markets: 2016-2025
  • Deep Learning Chipset Unit Shipments by Type, World Markets: 2016-2025
  • Deep Learning Chipset Revenue by Type, World Markets: 2016-2025
  • Deep Learning Chipset Revenue, Inference versus Trainig, World Markets: 2016-2025
  • Deep Learning Chipset Revenue by Compute Capacity, World Markets: 2016-2025
  • Deep Learning Chipset Revenue by Power Consumption, World Markets: 2016-2025
  • Deep Learning Chipset Average Selling Price by Type, World Markets: 2016-2025
  • Deep Learning Chipset Revenue by Market Sector, World Markets: 2016-2025
  • Deep Learning CPU Unit Shipments and Revenue, World Markets: 2016-2025
  • Deep Learning GPU Unit Shipments and Revenue, World Markets: 2016-2025
  • Deep Learning ASIC Unit Shipments and Revenue, World Markets: 2016-2025
  • Deep Learning FPGA Unit Shipments and Revenue, World Markets: 2016-2025
  • Deep Learning Other Chipset Unit Shipments and Revenue, World Markets: 2016-2025
  • Deep Learning Chipset Revenue by Industry, World Markets: 2016-2025
  • Deep Learning CPU Revenue by Region, World Markets: 2016-2025
  • Deep Learning CPU Revenue by Compute Capacity, World Markets: 2016-2025
  • Deep Learning CPU Revenue by Power Consumption, World Markets: 2016-2025
  • Deep Learning CPU Revenue, Inference versus Training, World Markets: 2016-2025
  • Deep Learning CPU Revenue by Market Sector, World Markets: 2016-2025
  • Deep Learning GPU Revenue by Region, World Markets: 2016-2025
  • Deep Learning GPU Revenue by Compute Capacity, World Markets: 2016-2025
  • Deep Learning GPU Revenue by Power Consumption, World Markets: 2016-2025
  • Deep Learning GPU Revenue, Inference versus Training, World Markets: 2016-2025
  • Deep Learning GPU Revenue by Market Sector, World Markets: 2016-2025
  • Deep Learning FPGA Revenue by Region, World Markets: 2016-2025
  • Deep Learning FPGA Revenue by Compute Capacity, World Markets: 2016-2025
  • Deep Learning FPGA Revenue by Power Consumption, World Markets: 2016-2025
  • Deep Learning FPGA Revenue, Inference versus Training, World Markets: 2016-2025
  • Deep Learning FPGA Revenue by Market Sector, World Markets: 2016-2025
  • Deep Learning Other Chipset Revenue by Region, World Markets: 2016-2025
  • Deep Learning Other Chipset Revenue by Compute Capacity, World Markets: 2016-2025
  • Deep Learning Other Chipset Revenue by Power Consumption, World Markets: 2016-2025
  • Deep Learning Other Chipset Revenue, Inference versus Training, World Markets: 2016-2025
  • Deep Learning Other Chipset Revenue by Market Sector, World Markets: 2016-2025
  • Deep Learning Chipset Enterprise Sector Revenue by Region, World Markets: 2016-2025
  • Deep Learning Chipset Enterprise Sector Year-on-Year Revenue Growth Rates, World Markets: 2016-2025
  • Deep Learning Chipset Enterprise Sector Revenue by Type, World Markets: 2016-2025
  • Deep Learning Chipset Enterprise Sector Revenue by Power Consumption, World Markets: 2016-2025
  • Deep Learning Chipset Enterprise Sector Revenue by Compute Capacity, World Markets: 2016-2025
  • Deep Learning Chipset Enterprise Sector Revenue, Inference versus Training, World Markets: 2016-2025
  • Deep Learning Chipset Consumer Sector Revenue by Region, World Markets: 2016-2025
  • Deep Learning Chipset Consumer Sector Year-on-Year Revenue Growth Rates, World Markets: 2016-2025
  • Deep Learning Chipset Consumer Sector Revenue by Type, World Markets: 2016-2025
  • Deep Learning Chipset Consumer Sector Revenue by Power Consumption, World Markets: 2016-2025
  • Deep Learning Chipset Consumer Sector Revenue by Compute Capacity, World Markets: 2016-2025
  • Deep Learning Chipset Consumer Sector Revenue, Inference versus Training, World Markets: 2016-2025
  • Deep Learning Chipset Defense Sector Revenue by Region, World Markets: 2016-2025
  • Deep Learning Chipset Defense Sector Year-on-Year Revenue Growth Rates, World Markets: 2016-2025
  • Deep Learning Chipset Defense Sector Revenue by Type, World Markets: 2016-2025
  • Deep Learning Chipset Defense Sector Revenue by Power Consumption, World Markets: 2016-2025
  • Deep Learning Chipset Defense Sector Revenue by Compute Capacity, World Markets: 2016-2025
  • Deep Learning Chipset Defense Sector Revenue, Inference versus Training, World Markets: 2016-2025

Figures

  • Evolution of Artificial Intelligence
  • Number of Companies Working with NVIDIA on Deep Learning
  • System Considerations When Choosing a Hardware Platform
  • Suitability of Different Chipsets for Training and Inference
  • CNN Algorithms Shown to Increase Accuracy of Vehicle Detection
  • Venn Diagram Describing Various Technologies in Deep Learning
  • Feedforward Network versus Recurrent Neural Network
  • A CNN Used for Image Recognition
  • Training versus Inference Illustrated
  • Comparison of GPU and CPU performance
  • Power Comparison for NVIDIA GPU Interface Engine
  • Improvement in Deep Learning Speed Algorithms over the last Few Years
  • Power Consumption of CPUs, GPUs, and FPGAs

Tables

  • Deep Learning Chipset Revenue by Region, World Markets: 2016-2025
  • Deep Learning Chipset Year-on-Year Revenue Growth Rates, World Markets: 2016-2025
  • Deep Learning Chipset Revenue by Type, World Markets: 2016-2025
  • Deep Learning Chipset Average Selling Price by Type, World Markets: 2016-2025
  • Deep Learning Chipset Unit Shipments by Type, World Markets: 2016-2025
  • Deep Learning Chipset Revenue by Market Sector, World Markets: 2016-2025
  • Deep Learning Chipset Revenue by Power Consumption, World Markets: 2016-2025
  • Deep Learning Chipset Revenue by Compute Capacity, World Markets: 2016-2025
  • Deep Learning Chipset Revenue, Inference versus Training, World Markets: 2016-2025
  • Deep Learning Chipset Revenue by Industry, World Markets: 2016-2025
  • Deep Learning CPU Revenue, World Markets: 2016-2025
  • Deep Learning CPU Average Selling Price, World Markets: 2016-2025
  • Deep Learning CPU Unit Shipments, World Markets: 2016-2025
  • Deep Learning CPU Revenue by Region, World Markets: 2016-2025
  • Deep Learning CPU Revenue by Compute Capacity, World Markets: 2016-2025
  • Deep Learning CPU Revenue by Power Consumption, World Markets: 2016-2025
  • Deep Learning CPU Revenue, Inference versus Training, World Markets: 2016-2025
  • Deep Learning CPU Revenue by Market Sector, World Markets: 2016-2025
  • Deep Learning GPU Revenue, World Markets: 2016-2025
  • Deep Learning GPU Average Selling Price, World Markets: 2016-2025
  • Deep Learning GPU Unit Shipments, World Markets: 2016-2025
  • Deep Learning GPU Revenue by Region, World Markets: 2016-2025
  • Deep Learning GPU Revenue by Compute Capacity, World Markets: 2016-2025
  • Deep Learning GPU Revenue by Power Consumption, World Markets: 2016-2025
  • Deep Learning GPU Revenue, Inference versus Training, World Markets: 2016-2025
  • Deep Learning GPU Revenue by Market Sector, World Markets: 2016-2025
  • Deep Learning FPGA Revenue, World Markets: 2016-2025
  • Deep Learning FPGA Average Selling Price, World Markets: 2016-2025
  • Deep Learning FPGA Unit Shipments, World Markets: 2016-2025
  • Deep Learning FPGA Revenue by Region, World Markets: 2016-2025
  • Deep Learning FPGA Revenue by Compute Capacity, World Markets: 2016-2025
  • Deep Learning FPGA Revenue by Power Consumption, World Markets: 2016-2025
  • Deep Learning FPGA Revenue, Inference versus Training, World Markets: 2016-2025
  • Deep Learning FPGA Revenue by Market Sector, World Markets: 2016-2025
  • Deep Learning ASIC Revenue, World Markets: 2016-2025
  • Deep Learning ASIC Average Selling Price, World Markets: 2016-2025
  • Deep Learning ASIC Unit Shipments, World Markets: 2016-2025
  • Deep Learning ASIC Revenue by Region, World Markets: 2016-2025
  • Deep Learning ASIC Revenue by Compute Capacity, World Markets: 2016-2025
  • Deep Learning ASIC Revenue by Power Consumption, World Markets: 2016-2025
  • Deep Learning ASIC Revenue, Inference versus Training, World Markets: 2016-2025
  • Deep Learning ASIC Revenue by Market Sector, World Markets: 2016-2025
  • Deep Learning Other Chipset Revenue, World Markets: 2016-2025
  • Deep Learning Other Chipset Average Selling Price, World Markets: 2016-2025
  • Deep Learning Other Chipset Unit Shipments, World Markets: 2016-2025
  • Deep Learning Other Chipset Revenue by Region, World Markets: 2016-2025
  • Deep Learning Other Chipset Revenue by Compute Capacity, World Markets: 2016-2025
  • Deep Learning Other Chipset Revenue by Power Consumption, World Markets: 2016-2025
  • Deep Learning Other Chipset Revenue, Inference versus Training, World Markets: 2016-2025
  • Deep Learning Other Chipset Revenue by Market Sector, World Markets: 2016-2025
  • Deep Learning Chipset Enterprise Sector Revenue by Region, World Markets: 2016-2025
  • Deep Learning Chipset Enterprise Sector Year-on-Year Revenue Growth Rates, World Markets: 2016-2025
  • Deep Learning Chipset Enterprise Sector Revenue by Type, World Markets: 2016-2025
  • Deep Learning Chipset Enterprise Sector Revenue by Power Consumption, World Markets: 2016-2025
  • Deep Learning Chipset Enterprise Sector Revenue by Compute Capacity, World Markets: 2016-2025
  • Deep Learning Chipset Enterprise Sector Revenue, Inference versus Training, World Markets: 2016-2025
  • Deep Learning Chipset Consumer Sector Revenue by Region, World Markets: 2016-2025
  • Deep Learning Chipset Consumer Sector Year-on-Year Revenue Growth Rates, World Markets: 2016-2025
  • Deep Learning Chipset Consumer Sector Revenue by Type, World Markets: 2016-2025
  • Deep Learning Chipset Consumer Sector Revenue by Power Consumption, World Markets: 2016-2025
  • Deep Learning Chipset Consumer Sector Revenue by Compute Capacity, World Markets: 2016-2025
  • Deep Learning Chipset Consumer Sector Revenue, Inference versus Training, World Markets: 2016-2025
  • Deep Learning Chipset Defense Sector Revenue by Region, World Markets: 2016-2025
  • Deep Learning Chipset Defense Sector Year-on-Year Revenue Growth Rates, World Markets: 2016-2025
  • Key Players in Different Deep Learning Chipsets
  • Comparison of Deep Learning Chipset Parameters
  • Overview of Deep Learning Frameworks
  • Percentage of Developers Frequently Using Various Deep Learning Frameworks
Back to Top