Deploying GPGPUs In Manufacturing
Deploying GPGPUs in ManufacturingA case study of software and CFD development in bicycle manufacturingACER CONFIDENTIAL
Overview Case overview Introduction to Velocite bicycles Introduction of Dr. Matthew Smith, key developer for the project Hardware provided Software used Usage scenarios CFD design and simulation Cloud-based file synchronization Code design and publishing Follow up Velocite bikes Matthew Smith Acer Inc. – US and TaiwanACER CONFIDENTIAL
Case overviewGoal Develop a new CFD solver and intelligent software allowingcomputer-based design and analysis for bicycle frame design Leverage the solver to create a more aerodynamic road bike stillcapable of holding up to the structural tasks of competition racingACER CONFIDENTIAL
Velocite Velocite’s slogan is ‘feel invincible’ and delivers through cutting edgebike design Always pushing the bounds of technology, the company looks todevelop the most aerodynamic bicycles in the worldAwards 2010 Microenterprise National Entrepreneurship Award winner2011 Taiwan Excellent Enterprise Quality Award winner2012 Taipei Cycle Show d&i and iF Award winner2012 Taiwan Golden Pin Design Award winner2012 ISPO BrandNew Performance Finalists2012 The women's entrepreneur competition - Elite Group – special commendationRecognition Member of World Federation of Sporting Goods Industry (WFSGI)ACER CONFIDENTIAL
Velocite – product performance leadershipVelocite partners with leading researchers and technology companies toensure its products are state-of-the-art Tainan Science Park Innovation and Incubation Center National Applied Research Laboratories (NARL) Acer Computers National Cheng Kung University (NCKU) Oxeon AB Others and pipelineACER CONFIDENTIAL
New Helios frame design Over 800,000 faces in the design Over 1000 hours of CFD Tested wheels and frame Various speeds and angles of attack Optimized for 6 degree angle of attack at 40 km/h Frame weight: 900g Best in class stiffness 150 N/mm torsional deflection UCI certified Available in Oct. 2013ACER CONFIDENTIAL
Prof. Matthew SmithEducation B. Eng. (Mechanical and Space Engineering), University of Queensland, Australia(1998-2001) M.Phil. (Mechanical Engineering), University of Queensland, Australia (20022003) Ph.D. (Mechanical Engineering), University of Queensland, Australia (2006-2008)Experience Simulation Engineer, Boeing Australia (2001-2002) Research Scholar, University of Queensland, Australia (2003-2003) Design and Production Engineer, Nordon Cylinders, Australia (2004-2006) Research Scholar, University of Queensland, Australia (2006-2007) Associate Researcher, National Center for High-performanceComputing, Taiwan (2008-2012) Assistant Professor, National Cheng-Kung University (Present)Expertise Rarefied Gas Dynamics (RGD) Computational Fluid Dynamics (CFD) High Performance Computing (HPC) General Purpose computing on Graphics Processing Units (GPGPU) CUDA and OpenCL Programming Augmented Reality (AR) Computer Vision (CV) Genetic Algorithms (GA) and OptimizationACER CONFIDENTIAL
Hardware usedAcer provided 2 servers for collaboration and developmentServer specifications: AT350 F2:2 x Intel Xeon E5-2670 processors2 x NVIDIA Tesla C20751 x SATA HDD for OS operation64 GB (8 x 8 GB) DDR3registered memory 1 x NVIDIA Quadro2000 5000 for 3D rendering ACER CONFIDENTIAL
Workstation solutionAcer’s AT350 F2 offers best of both worlds– PCIe controller embedded in CPU– Extreme memory density– Up to 3 GPUs in capacity– Best in class expandability– 80 PLUS powersupplies– Rack-mountable– Supercomputerin a box2 x Tesla C2075 (computation)1 x Quadro 2000 (rendering)ACER CONFIDENTIAL
Software usedCFD solver Qt for development UbuntuOne for file syncBazaar for version control Xflow used for final3D model simulation Solidworks for preliminarystress testingVelocite at ISC’12 with Acer collaborationACER CONFIDENTIAL
CFD solver development guidelines / goalsDevelopment of a CFD package must satisfy the following goals: Lightweight Multi-Platform Low development cost No complex function evaluations GeneralACER CONFIDENTIAL
Mathematical / physical theories appliedIn order to maintain both accuracy and simplicity within the solver,simplistic sum functions were employed rather than complextrigonometric or root-based algorithms Uniform Equilibrium Flux Method (UEFM)f (v) ΣWiH (ai ,bi) Quiet Direct Simulation (QDS) Qvf (v )dv ΣH(v)Qv Split Harten, Lax, van Leer (SHLL)ACER CONFIDENTIAL
Usage scenarios2 design locations Software development conducted in Tainan Testing conducted onsite in Kaohsiung Servers synced and interacted across theteams using UbuntuOne Other client PCs and tablets used to interactwith the simulation and edit software remotelyACER CONFIDENTIAL
Results achievedBike design Aerodynamic road bike with best in class design – good structuraldesign for stiffness while achieving superior aerodynamics UCI approved for use in professional racesSolver design and implementation CUDA solver was used to select the 2D airfoil profiles that were usedto create 3D shapes Two step selection process: benchmarking of airfoil shapes currently used on highperformance bicycles and genetic algorithm optimization of an airfoil shapeACER CONFIDENTIAL
Results achievedGreatest Challenges: Realizing high performance rendering speeds for large-scale three dimensionalobjects within the user interface.Integration of a cutting edge flow solver (high resolution, low stability and lowusability) with a easy-to-use interface.Real-time animated visualizations of large-scale data sets (millions of cells animatedat 10 fps).Mobile device integration!Performance versus a cluster: CUDA-powered CFD solver employed in the airfoil optimization exercises presented aspeedup of over 70x (using a single GPU) when compared with its CPU equivalent.Solver capable of using multiple (dual) GPU devices. Hence, performance equivalentto 140x.Advantage over a cluster of computers – no network communication bottlenecks orhardware required.ACER CONFIDENTIAL
Acer WorkstationsVeriton P seriesEnabling high performance RD for small businessesACER CONFIDENTIAL
Evolution of supercomputingProprietary to propagatedTITAN, Oak ridge National Labs (Cray), 2012Eniac, 1946Veriton P530 F2, Acer, 2013ACER CONFIDENTIAL
Acer Veriton WorkstationsManufacturing renderingSimulation – ANSYS, AltairHeavy video post processingLarge 3D design work ( 1000 objects)Veriton P130 F3Veriton P330 F2AutoCADAdobe Photoshop and 2D designBasic 3D creation, such as ProE, 3DSmaxACER CONFIDENTIALVeriton P530 F2
Veriton P130 F3 product summaryAcer’s Veriton P130 F3 workstation supporting the latest NVIDIA graphicscards along with Intel Xeon E3 processors delivers rock-solid performanceand easy expandability to ensure the success of your professional business.Intel Xeon E3 processors 4 DIMMs DDR3 ECC memoryNVIDIA Quadro 400NVIDIA Quadro 600NVIDIA Quadro 2000NVIDIA Quadro 4000ACER CONFIDENTIALUp to 4 HDDsOnboard RAID 0, 1, 5, 10
Veriton P330 F2 product summaryDesigned for users demanding an excellent combination of performance andexpandability, the Veriton P330 F3 is a best-of-class choice for bothcomputing and rendering capabilitiesIntel Xeon E5 processors 8 DIMMs DDR3 ECC memoryACER CONFIDENTIALNVIDIA Quadro 400NVIDIA Quadro 600NVIDIA Quadro 2000NVIDIA Quadro 4000NVIDIA Quardro K5000NVIDIA Quadro 6000Up to 4 HDDsOnboard RAID 0, 1, 5, 10
Veriton P530 F2 product summaryUp to 16 physical cores, and capable of supporting 3 GPUs, the Veriton P530F2 offers maximum compute and rendering performance to the mostdemanding designers and engineers2 x Intel Xeon E5 processors16 DIMMs DDR3 ECC memoryACER CONFIDENTIALNVIDIA Quadro 400NVIDIA Quadro 600NVIDIA Quadro 2000NVIDIA Quadro 4000NVIDIA Quardro K5000NVIDIA Quadro 6000Up to 4 HDDsOnboard RAID 0, 1, 5, 10
For more in-depth questionsAcer Inc. Acer US business inquiries:kamesh.kumar@acer.com Acer Workstation or HPC productsjoe.dutka@acer.com Also see us at booth #606Velocite Bicycles Please visit . Matthew Smith @ NCKU (Solver and/or interface inquiries) Email: msmith@mail.ncku.edu.twACER CONFIDENTIAL
Veriton P330 F2 product summary Designed for users demanding an excellent combination of performance and expandability, the Veriton P330 F3 is a best-of-class choice for both computing and rendering capabilities Intel Xeon E5 processors 8 DIMMs DDR3 ECC memory NVIDIA Quadro400 NVIDIA Quadro600 NVIDIA Quadro2000 NVIDIA Quadro4000 NVIDIA QuardroK5000
Citrix.com Deployment Guide Deploying Microsoft SharePoint 2016 with NetScaler 8 Deploying Microsoft SharePoint 2016 with NetScaler Deployment Guide After clicking OK, you will see the Basic Settings screen for the LB vserver. Here, you may change settings such as the session persi
Aug 01, 2016 · Oracle Database 12c Release 1 (12.1) on Red Hat Enterprise Linux 7. It is intended to provide a Red Hat Oracle reference architecture that focuses on the following tasks: Deploying Oracle Grid Infrastructure 12c Release 1 (12.1.0.2.0) Deploying Ora
7.2.1. Copying Prometheus metrics configuration to a custom resource 7.2.2. Deploying a Kafka cluster with Prometheus metrics configuration 7.3. VIEWING KAFKA METRICS AND DASHBOARDS IN OPENSHIFT 4 7.3.1. Deploying the Prometheus resources 7.3.2. Creating a Service Account for Grafana 7.3.3. Deploying Gra
The tutorial guides you through the following tasks: Downloading and installing your own Tomcat server. Coding and deploying a JSP on Tomcat. Coding and deploying a servlet on Tomcat. Coding and deploying a Web service using Tomcat and Apache Axis. The tutorial provides
Installing and Deploying Data Endpoint 1 Installing and Deploying Data Endpoint Clients Websense Data Endpoint is a solution for securing client workstations, laptops, and other endpoint machines from data loss when the machines are outside the corporate network and for identifying sensitive data on the clients themselves.
Deploying the BIG-IP LTM with IBM . Cognos Insight. Welcome to the F5 Deployment Guide for IBM Cognos Insight. This document provides guidance for deploying the BIG-IP Local Traffic Manager (LTM) with IBM Cognos. The BIG-IP LTM brings high availability, SSL offload, and TCP optimizations to IBM Cognos solutions.
activity monitoring. By deploying Guardium appliances to collect information from databases, your organization gains up-to-the-second insight into the activity happening at the application and data level. Now, by deploying the Database Security functionality within the BIG-IP system, you can correlate front-end information with database .
Description Logic Knowledge Base Exchange Elena Botoeva supervisor: Diego Calvanese PhD Final Examination April 10, 2014 Bolzano Elena Botoeva(FUB)Description Logic Knowledge Base Exchange1/33. Outline 1 Introduction 2 Summary of Work 3 Results 4 Technical Development Universal Solutions Universal UCQ-solutions UCQ-representations Elena Botoeva(FUB)Description Logic Knowledge Base Exchange2/33 .
