DOCSLIB.ORG

SLIDES [email protected]

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
SLIDES Yukidave@Ece.Neu.Edu 6/24/2013 WHAT IS THIS TALK ABOUT ? • Study of power and energy profile of different optimization ANALYZING OPTIMIZATION techniques used in heterogeneous applications TECHNIQUES FOR POWER EFFICIENCY ON HETEROGENEOUS • Evaluation of power/performance of such optimization techniques on heterogeneous applications such as FFT PLATFORMS Yash Ukidave and David Kaeli Department of Electrical and Computer Engineering Northeastern University, Boston, USA AsHES 2013 Boston, MA 20th May, 2013 1 | AsHES 2013 | May 2013 2 | AsHES 2013 | May 2013 TOPICS MOTIVATION Increasing use of the CPU-GPU Applications using OpenCL and power consumption of heterogeneous environment to accelerate data-parallel devices heterogeneous applications Fast Fourier Transforms (FFT) & evaluation methodology Thermal Design Power (TDP) of latest generation of GPUs used for Optimization techniques used for analysis heterogeneous compute RltfResults for power-performance o f FFT i mp lemen tati ons Understanding the effects of software design methods contributing to power Analysis of power-performance of different optimization techniques consumption Energy profile of different optimization techniques Power and Energy aspect of different optimization techniques for Conclusion heterogeneous platforms Future work 3 | AsHES 2013 | May 2013 4 | AsHES 2013 | May 2013 1 6/24/2013 FAST FOURIER TRANSFORM (FFT) x(0) X(0) FFT IMPLEMENTATIONS FFT is an algorithm to compute Discrete Fourier • MR-SC FFT : Multi-Radix single FFT Kernel Memory Twiddle Transforms (DFT) kernel call, based on the FFT Implementation Calls Access factor x(1) X(1) Patterns Computation -1 implementation in AMD SDK. Reduces time complexity to O(nlogn) from O(n2) MR-SC Global & Kernel • MR-MC FFT : Based on Cooley- FFT classified as Decimation in Time(DIT) or Single Local Compile-Time Tukey algorithm, uses Multiple Decimation in Frequency (DIF) Memory kernel calls and Multiple Radix combinations for compute DIT : Operates on odd and even components of Stockham FFT signal Kernel • Stockham FFT : Based on the DIF : Operates on two halves of the signal Apple FFT Global Run-Time Stockham algorithm for FFT. Single Multiple Memory Butterfly structure is a major component of FFT of a Radix and Single Kernel call MR-MC FFT given Radix computation Each FFT work item performs computes butterfly on • Apple FFT : A Multiple Kernel call given data points based FFT provided by Apple Inc. using OpenCL 5 | AsHES 2013 | May 2013 6 | AsHES 2013 | May 2013 PLATFORMS FOR EVALUATION OPTIMIZATION TECHNIQUES CLASSIFIED IN SETS Shared Memory APUs Discrete GPUs Set S0 Device Features Intel AMD Device Features Nvidia AMD No modifications Core i7 Fusion A8 GTX 680 Radeon HD 3300 APU 7770 ( “out-of-box”) Device Generation Ivy Evergreen Device Generation Kepler Southern performance Bridge Islands Compute 16 5 Compute 810 Units(()CU) Units(()CU) Processing 416 Processing 192 64 Set S1 Elements(PE)/CU Elements(PE)/CU Set S2 Set S3 Coalesced Global Data Transformation Local Memory TDP(Watts) 70 100 TDP(Watts) 195 80 Memory Accesses Usage float Æ float2 Memory 26 17 Memory 192 72 Loop Unrolling Stage overlapping Bandwidth(GB/s) Bandwidth(GB/s) float Æ float4 for stage-based compute Register File 64 64 Register File 256 256 float Æ float8 Per CU(KB) Per CU(KB) 7 | AsHES 2013 | May 2013 8 | AsHES 2013 | May 2013 2 6/24/2013 EXECUTION PERFORMANCE OF FFT POWER PERFORMANCE OF NON-OPTIMIZED FFT • Performance is compared to the baseline on each platform High power consumption of Nvidia GPUs affect their power efficiency over AMD • Evaluation is done for three data set sizes 64K, 1M and 2M data points GPUs Execution Performance on GPUs Execution Performance on APUs Low compute capability of APUs exhibit decrease in performance over GPUs Power Performance on GPUs Power Performance on APUs 9 | AsHES 2013 | May 2013 10 | AsHES 2013 | May 2013 ANALYSIS OF MEMORY BASED OPTIMIZATIONS ANALYSIS OF MEMORY BASED OPTIMIZATIONS S1 & S2 optimizations use coalesced memory accesses in FFT kernels Improved throughput causes increase in FFTs are modified to perform coalesced memory accesses and loop structures execution performance are unrolled Coalesced accesses are Improved throughput on devices leads to increase in performance handled using specialized Average improvement in throughput is 2X on discrete GPUs hardware paths Coalesced accesses increase power consumption Average increase in power is 32% 11 | AsHES 2013 | May 2013 12 | AsHES 2013 | May 2013 3 6/24/2013 ANALYSIS OF MEMORY BASED OPTIMIZATIONS EFFECTS OF LOCAL MEMORY USAGE AND STAGE-OVERLAPPING S2 data transform evaluation on Nvidia GPU S2 optimizations use data • S3 optimizations use local transformation memory and stage-overlapping Data transforms in OpenCL • Stages of multi-kernel allow contiguous memory implementation are overlapped access in single kernel call This increases coalesced • MRMC and Apple FFT are accesses for the kernel modified for stage overlap Power consumption increases • Kernel call overhead is due to coalesced access avoided and coalesced access are also used Per-workitem compute increases due to increase in • Large performance gains are input data access observed on GPUs and APUs 13 | AsHES 2013 | May 2013 14 | AsHES 2013 | May 2013 ARCHITECTURAL FACTORS CONTRIBUTING TO POWER ENERGY PROFILE OF OPTIMIZATION TECHNIQUES S3 optimizations are evaluated Energy relates power for this analysis consumption and execution performance directly Memory stalls and ALU utilization is evaluated for power Power consumption can consumption increase due to increase in performance Memoryyy unit stalls directly affects power consumption of Energy consumption exposes GPUs and APUs this trade-off Number of in-flight memory Variation in energy across accesses can cause memory- optimizations is 11% on average unit stalls S2 optimizations show 13% ALU utilization does not directly increase in energy over other affect power consumption optimization sets 15 | AsHES 2013 | May 2013 16 | AsHES 2013 | May 2013 4 6/24/2013 RESULTS SUMMARY RESULTS SUMMARY 9 Highly Efficient 8 Less Efficient + Moderately Efficient More than 40% Less than 10% 10-40% improvement GPUs APUs improvement improvement • S1 & S2 improve performance • S1 & S2 are not power efficient Power Performance Power Performance efficiencyyp with cost for power Optimization Techniques Efficiency Efficiency (Gflops/Watts) consumption • Local memory improves compute efficiency and power efficiency • Local memory in S3 improve GPUs APUs GPUs APUs GPUs APUs power-performance • Stage overlapping increases load Coalesced Memory Access 8899+ + on resources and increases power Loop Unrolling 8899+ + • S3 are compute and power efficient consumption Data Transformation 88998 8 Local Memory Usage 99999 9 Stage Overlapping 9+999 + 17 | AsHES 2013 | May 2013 18 | AsHES 2013 | May 2013 CONCLUSION FUTURE WORK Analyzed different optimization techniques for their power-performance on once Analyze power consumption on SoC (System-on-Chip) devices with GPUs, such class of applications as TI OMAP4, Samsung Exynos, Qualcomm Snapdragon Study helps developer identify potential of power-aware kernel development Power-performance analysis to multi-GPU environments such as clusters Optimizations related to coalesced memory accesses exhibit increase in power Study of microarchitectural features responsible for power consumption on GPUs consumption on GPUs and APUs Extend study to different heterogeneous multi-core devices such as Adapteva Local Memory utilization is observed as the most power efficient optimization Ephiphany Processor, Tilera TilePro processors technique Power increment due to performance improvement is captured effectively in the energy profile 19 | AsHES 2013 | May 2013 20 | AsHES 2013 | May 2013 5 6/24/2013 THANK YOU ! QUESTIONS ? COMMENTS ? Yash Ukidave EXTRA SLIDES [email protected] 21 | AsHES 2013 | May 2013 22 | AsHES 2013 | May 2013 TITLE PAGE PICTURE FFT IMPLEMENTATIONS WITH SINGLE KERNEL CALL MR-SC FFT : FFT implementation based on AMD SDK. Multiple Radix sizes used Stockham FFT : Stockham algorithm for FFT computation. Single Radix computation FFT Source: http://commons.wikimedia.org/wiki/File:Mona_Lisa_bw_square.jpeg 23 | AsHES 2013 | May 2013 24 | AsHES 2013 | May 2013 6 6/24/2013 FFT IMPLEMENTATIONS WITH MULTIPLE KERNEL CALLS POWER MEASUREMENT SETUP Power measurement on GPU: MR-MC FFT : Multiple kernel calls and Multiple Radix combinations for compute Discrete GPUs were isolated from the system for power measurement Apple FFT : Multiple Kernel call based FFT by Apple Inc. using OpenCL A dedicated external Power supply is used for Discrete GPUs Power consumption of GPU is measured by profiling the External PSU Power measurement on GPUs Power Measurement for APU: Power supply for APUs is measured using a similar power meter System power is measured to record power consumption on APUs Graphic processor of APUs cannot be completely isolated off the host CPU device Power measurement on APUs 25 | AsHES 2013 | May 2013 26 | AsHES 2013 | May 2013 7.
Load more

Recommended publications
  • A 1024-Core 70GFLOPS/W Floating Point Manycore Microprocessor
    A 1024-core 70GFLOPS/W Floating Point Manycore Microprocessor Andreas Olofsson, Roman Trogan, Oleg Raikhman Adapteva, Lexington, MA The Past, Present, & Future of Computing SIMD MIMD PE PE PE PE MINI MINI MINI CPU CPU CPU PE PE PE PE MINI MINI MINI CPU CPU CPU PE PE PE PE MINI MINI MINI CPU CPU CPU MINI MINI MINI BIG BIG CPU CPU CPU CPU CPU BIG BIG BIG BIG CPU CPU CPU CPU PAST PRESENT FUTURE 2 Adapteva’s Manycore Architecture C/C++ Programmable Incredibly Scalable 70 GFLOPS/W 3 Routing Architecture 4 E64G400 Specifications (Jan-2012) • 64-Core Microprocessor • 100 GFLOPS performance • 800 MHz Operation • 8GB/sec IO bandwidth • 1.6 TB/sec on chip memory BW • 0.8 TB/sec network on chip BW • 64 Billion Messages/sec IO Pads Core • 2 Watt total chip power • 2MB on chip memory Link Logic • 10 mm2 total silicon area • 324 ball 15x15mm plastic BGA 5 Lab Measurements 80 Energy Efficiency 70 60 50 GFLOPS/W 40 30 20 10 0 0 200 400 600 800 1000 1200 MHz ENERGY EFFICIENCY ENERGY EFFICIENCY (28nm) 6 Epiphany Performance Scaling 16,384 G 4,096 F 1,024 L 256 O 64 4096 P 1024 S 16 256 64 4 16 1 # Cores On‐demand scaling from 0.25W to 64 Watt 7 Hold on...the title said 1024 cores! • We can build it any time! • Waiting for customer • LEGO approach to design • No global timinga paths • Guaranteed by design • Generate any array in 1 day • ~130 mm2 silicon area 1024 Cores 1Core 8 What about 64-bit Floating Point? Single Precision Double Precision 2 FLOPS/CYCLE 2 FLOPS/CYCLE 64KB SRAM 64KB SRAM 0.215mm^2 0.237mm^2 700MHz 600MHz 9 Epiphany Latency Specifications
    [Show full text]
  • Comparison of 116 Open Spec, Hacker Friendly Single Board Computers -- June 2018
    Comparison of 116 Open Spec, Hacker Friendly Single Board Computers -- June 2018 Click on the product names to get more product information. In most cases these links go to LinuxGizmos.com articles with detailed product descriptions plus market analysis. HDMI or DP- USB Product Price ($) Vendor Processor Cores 3D GPU MCU RAM Storage LAN Wireless out ports Expansion OSes 86Duino Zero / Zero Plus 39, 54 DMP Vortex86EX 1x x86 @ 300MHz no no2 128MB no3 Fast no4 no5 1 headers Linux Opt. 4GB eMMC; A20-OLinuXino-Lime2 53 or 65 Olimex Allwinner A20 2x A7 @ 1GHz Mali-400 no 1GB Fast no yes 3 other Linux, Android SATA A20-OLinuXino-Micro 65 or 77 Olimex Allwinner A20 2x A7 @ 1GHz Mali-400 no 1GB opt. 4GB NAND Fast no yes 3 other Linux, Android Debian Linux A33-OLinuXino 42 or 52 Olimex Allwinner A33 4x A7 @ 1.2GHz Mali-400 no 1GB opt. 4GB NAND no no no 1 dual 40-pin 3.4.39, Android 4.4 4GB (opt. 16GB A64-OLinuXino 47 to 88 Olimex Allwinner A64 4x A53 @ 1.2GHz Mali-400 MP2 no 1GB GbE WiFi, BT yes 1 40-pin custom Linux eMMC) Banana Pi BPI-M2 Berry 36 SinoVoip Allwinner V40 4x A7 Mali-400 MP2 no 1GB SATA GbE WiFi, BT yes 4 Pi 40 Linux, Android 8GB eMMC (opt. up Banana Pi BPI-M2 Magic 21 SinoVoip Allwinner A33 4x A7 Mali-400 MP2 no 512MB no Wifi, BT no 2 Pi 40 Linux, Android to 64GB) 8GB to 64GB eMMC; Banana Pi BPI-M2 Ultra 56 SinoVoip Allwinner R40 4x A7 Mali-400 MP2 no 2GB GbE WiFi, BT yes 4 Pi 40 Linux, Android SATA Banana Pi BPI-M2 Zero 21 SinoVoip Allwinner H2+ 4x A7 @ 1.2GHz Mali-400 MP2 no 512MB no no WiFi, BT yes 1 Pi 40 Linux, Android Banana
    [Show full text]
  • Survey and Benchmarking of Machine Learning Accelerators
    1 Survey and Benchmarking of Machine Learning Accelerators Albert Reuther, Peter Michaleas, Michael Jones, Vijay Gadepally, Siddharth Samsi, and Jeremy Kepner MIT Lincoln Laboratory Supercomputing Center Lexington, MA, USA freuther,pmichaleas,michael.jones,vijayg,sid,[email protected] Abstract—Advances in multicore processors and accelerators components play a major role in the success or failure of an have opened the flood gates to greater exploration and application AI system. of machine learning techniques to a variety of applications. These advances, along with breakdowns of several trends including Moore’s Law, have prompted an explosion of processors and accelerators that promise even greater computational and ma- chine learning capabilities. These processors and accelerators are coming in many forms, from CPUs and GPUs to ASICs, FPGAs, and dataflow accelerators. This paper surveys the current state of these processors and accelerators that have been publicly announced with performance and power consumption numbers. The performance and power values are plotted on a scatter graph and a number of dimensions and observations from the trends on this plot are discussed and analyzed. For instance, there are interesting trends in the plot regarding power consumption, numerical precision, and inference versus training. We then select and benchmark two commercially- available low size, weight, and power (SWaP) accelerators as these processors are the most interesting for embedded and Fig. 1. Canonical AI architecture consists of sensors, data conditioning, mobile machine learning inference applications that are most algorithms, modern computing, robust AI, human-machine teaming, and users (missions). Each step is critical in developing end-to-end AI applications and applicable to the DoD and other SWaP constrained users.
    [Show full text]
  • SPORK: a Summarization Pipeline for Online Repositories of Knowledge
    SPORK: A SUMMARIZATION PIPELINE FOR ONLINE REPOSITORIES OF KNOWLEDGE A Thesis presented to the Faculty of California Polytechnic State University San Luis Obispo In Partial Fulfillment of the Requirements for the Degree Master of Science in Computer Science by Steffen Lyngbaek June 2013 c 2013 Steffen Lyngbaek ALL RIGHTS RESERVED ii COMMITTEE MEMBERSHIP TITLE: SPORK: A Summarization Pipeline for Online Repositories of Knowledge AUTHOR: Steffen Lyngbaek DATE SUBMITTED: June 2013 COMMITTEE CHAIR: Professor Alexander Dekhtyar, Ph.D., De- parment of Computer Science COMMITTEE MEMBER: Professor Franz Kurfess, Ph.D., Depar- ment of Computer Science COMMITTEE MEMBER: Professor Foaad Khosmood, Ph.D., Depar- ment of Computer Science iii Abstract SPORK: A Summarization Pipeline for Online Repositories of Knowledge Steffen Lyngbaek The web 2.0 era has ushered an unprecedented amount of interactivity on the Internet resulting in a flood of user-generated content. This content is of- ten unstructured and comes in the form of blog posts and comment discussions. Users can no longer keep up with the amount of content available, which causes developers to start relying on natural language techniques to help mitigate the problem. Although many natural language processing techniques have been em- ployed for years, automatic text summarization, in particular, has recently gained traction. This research proposes a graph-based, extractive text summarization system called SPORK (Summarization Pipeline for Online Repositories of Knowl- edge). The goal of SPORK is to be able to identify important key topics presented in multi-document texts, such as online comment threads. While most other automatic summarization systems simply focus on finding the top sentences rep- resented in the text, SPORK separates the text into clusters, and identifies dif- ferent topics and opinions presented in the text.
    [Show full text]
  • Master's Thesis: Adaptive Core Assignment for Adapteva Epiphany
    Adaptive core assignment for Adapteva Epiphany Master of Science Thesis in Embedded Electronic System Design Erik Alveflo Chalmers University of Technology Department of Computer Science and Engineering G¨oteborg, Sweden 2015 The Author grants to Chalmers University of Technology and University of Gothenburg the non-exclusive right to publish the Work electronically and in a non-commercial pur- pose make it accessible on the Internet. The Author warrants that he/she is the author to the Work, and warrants that the Work does not contain text, pictures or other mate- rial that violates copyright law. The Author shall, when transferring the rights of the Work to a third party (for example a publisher or a company), acknowledge the third party about this agreement. If the Author has signed a copyright agreement with a third party regarding the Work, the Author warrants hereby that he/she has obtained any necessary permission from this third party to let Chalmers University of Technology and University of Gothenburg store the Work electronically and make it accessible on the Internet. Adaptive core assignment for Adapteva Epiphany Erik Alveflo, c Erik Alveflo, 2015. Examiner: Per Larsson-Edefors Chalmers University of Technology Department of Computer Science and Engineering SE-412 96 G¨oteborg Sweden Telephone + 46 (0)31-772 1000 Department of Computer Science and Engineering G¨oteborg, Sweden 2015 Adaptive core assignment for Adapteva Epiphany Erik Alveflo Department of Computer Science and Engineering Chalmers University of Technology Abstract The number of cores in many-core processors is ever increasing, and so is the number of defects due to manufacturing variations and wear-out mechanisms.
    [Show full text]
  • Program & Exhibits Guide
    FROM CHIPS TO SYSTEMS – LEARN TODAY, CREATE TOMORROW CONFERENCE PROGRAM & EXHIBITS GUIDE JUNE 24-28, 2018 | SAN FRANCISCO, CA | MOSCONE CENTER WEST Mark You Calendar! DAC IS IN LAS VEGAS IN 2019! MACHINE IP LEARNING ESS & AUTO DESIGN SECURITY EDA IoT FROM CHIPS TO SYSTEMS – LEARN TODAY, CREATE TOMORROW JUNE 2-6, 2019 LAS VEGAS CONVENTION CENTER LAS VEGAS, NV DAC.COM DAC.COM #55DAC GET THE DAC APP! Fusion Technology Transforms DOWNLOAD FOR FREE! the RTL-to-GDSII Flow GET THE LATEST INFORMATION • Fusion of Best-in-Class Optimization and Industry-golden Signoff Tools RIGHT WHEN YOU NEED IT. • Unique Fusion Data Model for Both Logical and Physical Representation DAC.COM • Best Full-flow Quality-of-Results and Fastest Time-to-Results MONDAY SPECIAL EVENT: RTL-to-GDSII Fusion Technology • Search the Lunch at the Marriott Technical Program • Find Exhibitors www.synopsys.com/fusion • Create Your Personalized Schedule Visit DAC.com for more details and to download the FREE app! GENERAL CHAIR’S WELCOME Dear Colleagues, be able to visit over 175 exhibitors and our popular DAC Welcome to the 55th Design Automation Pavilion. #55DAC’s exhibition halls bring attendees several Conference! new areas/activities: It is great to have you join us in San • Design Infrastructure Alley is for professionals Francisco, one of the most beautiful who manage the HW and SW products and services cities in the world and now an information required by design teams. It houses a dedicated technology capital (it’s also the city that Design-on-Cloud Pavilion featuring presentations my son is named after).
    [Show full text]
  • Proyecto Fin De Grado
    ESCUELA TÉCNICA SUPERIOR DE INGENIERÍA Y SISTEMAS DE TELECOMUNICACIÓN PROYECTO FIN DE GRADO TÍTULO: Despliegue de Liota (Little IoT Agent) en Raspberry Pi AUTOR: Ricardo Amador Pérez TITULACIÓN: Ingeniería Telemática TUTOR (o Director en su caso): Antonio da Silva Fariña DEPARTAMENTO: Departamento de Ingeniería Telemática y Electrónica VºBº Miembros del Tribunal Calificador: PRESIDENTE: David Luengo García VOCAL: Antonio da Silva Fariña SECRETARIO: Ana Belén García Hernando Fecha de lectura: Calificación: El Secretario, Despliegue de Liota (Little IoT Agent) en Raspberry Pi Quizás de todas las líneas que he escrito para este proyecto, estas sean a la vez las más fáciles y las más difíciles de todas. Fáciles porque podría doblar la longitud de este proyecto solo agradeciendo a mis padres la infinita paciencia que han tenido conmigo, el apoyo que me han dado siempre, y el esfuerzo que han hecho para que estas líneas se hagan realidad. Por todo ello y mil cosas más, gracias. Mamá, papá, lo he conseguido. Fáciles porque sin mi tutor Antonio, este proyecto tampoco sería una realidad, no solo por su propia labor de tutor, si no porque literalmente sin su ayuda no se hubiera entregado a tiempo y funcionando. Después de esto Antonio, voy a tener que dejarme ganar algún combate en kenpo como agradecimiento. Fáciles porque, sí melones os toca a vosotros, Alex, Alfonso, Manu, Sama, habéis sido mi apoyo más grande en los momentos más difíciles y oscuros, y mis mejores compañeros en los momentos de felicidad. Amigos de Kulturales, los hermanos Baños por empujarme a mejorar, Pablo por ser un ejemplo a seguir, Chou, por ser de los mejores profesores y amigos que he tenido jamás.
    [Show full text]
  • Experiences Using Tegra K1 and X1 for Highly Energy Efficient Computing
    Experiences Using Tegra K1 and X1 for Highly Energy Efficient Computing Gaurav Mitra Andrew Haigh Luke Angove Anish Varghese Eric McCreath Alistair P. Rendell Research School of Computer Science Australian National University Canberra, Australia April 07, 2016 Introduction & Background Overview 1 Introduction & Background 2 Power Measurement Environment 3 Experimental Platforms 4 Approach 5 Results & Analysis 6 Conclusion Mitra et. al. (ANU) GTC 2016, San Francisco April 07, 2016 2 / 20 Introduction & Background Use of low-powered SoCs for HPC Nvidia Jetson TK1: ARM + GPU SoC Nvidia Jetson TX1: ARM + GPU SoC TI Keystone II: ARM + DSP SoC Adapteva Parallella: ARM + 64-core NoC TI BeagleBoard: ARM + DSP SoC Terasic DE1: ARM + FPGA SoC Rockchip Firefly: ARM + GPU SoC Freescale Wandboard: ARM + GPU SoC Cubieboard4: ARM + GPU SoC http://cs.anu.edu.au/systems Mitra et. al. (ANU) GTC 2016, San Francisco April 07, 2016 3 / 20 Introduction & Background Use of low-powered SoCs for HPC In order for SoC processors to be considered viable exascale building blocks, important factors to explore include: Absolute performance Balancing use of different on-chip devices Understanding the performance-energy trade-off Mitra et. al. (ANU) GTC 2016, San Francisco April 07, 2016 4 / 20 Introduction & Background Contributions Environment for monitoring and collecting high resolution power measurements for SoC systems Understanding the benefits of exploiting both the host CPU and accelerator GPU cores simultaneously for critical HPC kernels Performance and energy comparisons
    [Show full text]
  • AI-Optimized Chipsets
    AI-Optimized Chipsets Part III: Key Opportunities & Trends Aug 2018 An Introduction Previously in Part I, we reviewed the ADAC loop and key factors driving innovation for AI- optimized chipsets. In Part II, we review the shift in performance focus computing from general application neural nets and how this is driving demand for high performance computing. To this end, some startups are adopting alternative, novel approaches and this is expected to pave the way for other AI-optimized chipsets. In this instalment, we review the training and inference chipset markets, assess the dominance of tech giants, as well as the startups adopting cloud-first or edge-first approaches to AI-optimized chipsets. The training chipset market is dominated by 5 firms, while the inference chipset market is more diverse with >40 players Training Chipsets Inference Chipsets No. of Companies • 5 (i.e. Nvidia, Intel, Xilinx, AMD, Google) • > 40 Differentiators • Computation Performance • Power Efficiency • Usability • System Latency • Innovation Road Map • Cost • Computation Performance Barriers to Entry • R&D Intensity • Manufacturing Scale Economies • Developer Support • High Switching Costs (End User) • Size of End Markets • Regulatory Requirements • Distribution ASP • USD 2,000 - USD 20,000 • USD 80 - USD 10,000 Source: UBS • The training chipset market is dominated by 5 firms which have developed massively parallel architectures, well-suited for deep learning algorithms. NVIDIA is the most prominent with its GPU technology stack. • The early leaders in this market are likely to maintain their lead, but the inference market is large and easily accessible to all, including tech giants and startups. Given the significant market for inference chipsets, many defensible market niches based on high-system speed, low power and/or low Total Cost of Ownership (TCO) products are likely to emerge.
    [Show full text]
  • November 2–5, 2014 Asilomar Hotel and Conference Grounds
    Monterey, CA 93943 CA Monterey, 8236 Box P.O. Corp. SS&C Conf. FORTY-EIGHTH ASILOMAR CONFERENCE ON SIGNALS, SYSTEMS AND COMPUTERS November 2–5, 2014 Asilomar Hotel and Conference Grounds Technical Co-sponsor FORTY-EIGHTH Welcome from the General Chairman ASILOMAR CONFERENCE ON Prof. Roger Woods SIGNALS, SYSTEMS & COMPUTERS Queen’s University Belfast, UK Welcome to the 48th Asilomar Conference on Signals, Systems, and Computers! I have had a long involvement with the Conference since my first publication in 1997 when I was immediately struck by the unique nature of the Asilomar conference environment. The picturesque sand dunes and warm sunshine provide a wonderful backdrop to a conference that allows easy access to, and Technical Co-sponsor interaction with key researchers. Understandably, over the years, I have needed little persuasion to attend. There will never be a better opportunity to capture the attention of a key researcher in your area IEEE SIGNAL PROCESSING SOCIETY of expertise than at Asilomar! The technical program was crafted expertly by the Technical Program Chair, Geert Leus, and his team of Technical Area Chairs: Shengli Zhou, Zhengdao Wang, Bhaskar Rao, Michael Rabbat, Zhi Tian, Visa Koivunen, Selin Aviyente, Jorn Janneck, Mohsin Jamali, and Matt McKay. I would like to thank Geert and his team for assembling a high quality program with 439 accepted papers and 164 invited papers. The student paper contest this year has been chaired by Joe Cavallaro and he has selected a total of 11 CONFERENCE COMMITTEE submissions. The student finalists will present poster presentations to the judges on Sunday afternoon and of course, everyone is General Chair Publicity Chair welcome to attend.
    [Show full text]
  • Electronics Tinkering Books
    Books about Home and Classroom Electronics Tinkering, Hacking, Programming, and having fun with Raspberry Pi, Arduino, Intel Edison, Intel Galileo, and the like Collection of References edited by Stanislav Sýkora Permalink via DOI: 10.3247/SL5Refs15.003 Stan's LIBRARY and its Electronics Section Extra Byte | Stan's HUB Free online electronics texts Forward a missing book reference Site Plan & SEARCH This almost complete compilation includes titles yet to be released (they have a month specified in the release date). The entries are sorted by publication year and the first Author. Green-color titles indicate educational texts. You can download a PDF version of this document for off-line use. But keep coming back, the list is growing! Many of the books are available from Amazon. Entering Amazon from here helps this site at no cost to you. F Other Lists: Popular Science F Mathematics F Physics F Chemistry Visitor # Patents+IP F Electronics | DSP | Tinkering F Computing Spintronics F Materials ADVERTISE with us WWW issues F Instruments / Measurements Quantum Computing F NMR | ESR | MRI F Spectroscopy Extra Byte Hint: the F symbols above, where present, are links to free online texts (books, courses, theses, ...) Advance notices (years ≥ 2015). At page bottom, Related Works, HARDWARE, and LINKS: Link Directories: SCIENCE | Edu+Fun 1. Benjamin Alex, MATH | COMPUTING Raspberry Pi 2: Beginners User Manual, PHYSICS | CHEMISTRY Independent 2015. Kindle >>. NMR-MRI-ESR-NQR 2. Borchers Jan, ELECTRONICS Arduino in a Nutshell, PATENTS+IP FREE, 2015. Online >>. WWW stuff 3. Bradt Paul, Bradt David, Opaskar Joanna, Arduino Force, Pressure, Friction, and Acceleration Science Fair Projects, Other resources: CreateSpace Independent 2015.
    [Show full text]
  • Analysis of Task Scheduling for Multi-Core Embedded Systems
    Analysis of task scheduling for multi-core embedded systems Analys av schemaläggning för multikärniga inbyggda system JOSÉ LUIS GONZÁLEZ-CONDE PÉREZ, MASTER THESIS Supervisor: Examiner: De-Jiu Chen, KTH Martin Törngren, KTH Detlef Scholle, XDIN AB Barbro Claesson, XDIN AB MMK 2013:49 MDA 462 Acknowledgements I would like to thank my supervisors Detlef Scholle and Barbro Claesson for giving me the opportunity of doing the Master thesis at XDIN. I appreciate the kindness of Barbro chatting with me in Spanish and the support of Detlef no matter how much time it was required. I want to thank Sebastian, David and the other people at XDIN for the nice environment I lived in during these 20 weeks. I would like to thank the support and guidance of my supervisor at KTH DJ Chen and the help of my examiner Martin Törngren in the last stage of the thesis. I want to thank very much the other thesis colleagues at XDIN Joanna, Cheuk, Amir, Robin and Tobias. You have done this experience a lot more enriching. I would like to say merci! to my friends from Tyresö Benoit, Perrine, Simon, Audrey, Pierre, Marie-Line, Roberto, Alberto, Iván, Vincent, Olivier, Achour, Maxime, Si- mon, Emilie, Adelie, Siim and all the others. I have had great memories with you during the first year at KTH. I thank Osman and Tarek for this year in Midsom- markransen. I thank all the professors and staff from the Mechatronics department Mike, Bengt, Chen, Kalle, Jad and the others for making this programme possible, es- pecially Martin Edin Grimheden for his commitment with the students.
    [Show full text]