Journée Mésochallenges 2015 SysFera and ROMEO Make Large-Scale CFD Simulations Only 3 Clicks Away



Similar documents
ME6130 An introduction to CFD 1-1

HPC Cluster Decisions and ANSYS Configuration Best Practices. Diana Collier Lead Systems Support Specialist Houston UGM May 2014

Enterprise HPC & Cloud Computing for Engineering Simulation. Barbara Hutchings Director, Strategic Partnerships ANSYS, Inc.

Building a Top500-class Supercomputing Cluster at LNS-BUAP

Mississippi State University High Performance Computing Collaboratory Brief Overview. Trey Breckenridge Director, HPC

High Performance Computing in CST STUDIO SUITE

Accelerating CFD using OpenFOAM with GPUs

Cluster Scalability of ANSYS FLUENT 12 for a Large Aerodynamics Case on the Darwin Supercomputer

Recent Advances in HPC for Structural Mechanics Simulations

A GPU COMPUTING PLATFORM (SAGA) AND A CFD CODE ON GPU FOR AEROSPACE APPLICATIONS

Introduction to CFD Analysis

HPC Cloud Computing Guide PENGUIN ( ) HPC

Introduction to CFD Analysis

Numerical Calculation of Laminar Flame Propagation with Parallelism Assignment ZERO, CS 267, UC Berkeley, Spring 2015

Pedraforca: ARM + GPU prototype

STORAGE HIGH SPEED INTERCONNECTS HIGH PERFORMANCE COMPUTING VISUALISATION GPU COMPUTING

Express Introductory Training in ANSYS Fluent Lecture 1 Introduction to the CFD Methodology

Accelerating Simulation & Analysis with Hybrid GPU Parallelization and Cloud Computing

Very special thanks to Wolfgang Gentzsch and Burak Yenier for making the UberCloud HPC Experiment possible.

HETEROGENEOUS HPC, ARCHITECTURE OPTIMIZATION, AND NVLINK

LBM BASED FLOW SIMULATION USING GPU COMPUTING PROCESSOR

Kriterien für ein PetaFlop System

Customer Training Material. Lecture 2. Introduction to. Methodology ANSYS FLUENT. ANSYS, Inc. Proprietary 2010 ANSYS, Inc. All rights reserved.

GTC Presentation March 19, Copyright 2012 Penguin Computing, Inc. All rights reserved

CALMIP a Computing Meso-center for middle HPC academis Users

TSUBAME-KFC : a Modern Liquid Submersion Cooling Prototype Towards Exascale

Remote Visualization and Collaborative Design for CAE Applications

1 Bull, 2011 Bull Extreme Computing

MEETING THE CHALLENGES OF COMPLEXITY AND SCALE FOR MANUFACTURING WORKFLOWS

Accelerating Innovation with Self- Service HPC

SGI HPC Systems Help Fuel Manufacturing Rebirth

ECLIPSE Best Practices Performance, Productivity, Efficiency. March 2009

Jezelf Groen Rekenen met Supercomputers

GPU System Architecture. Alan Gray EPCC The University of Edinburgh

Overview of HPC systems and software available within

OpenMP Programming on ScaleMP

THE CFD SIMULATION OF THE FLOW AROUND THE AIRCRAFT USING OPENFOAM AND ANSA

CCTech TM. ICEM-CFD & FLUENT Software Training. Course Brochure. Simulation is The Future

ANSYS Computing Platform Support. July 2013

High-Performance Computing and Big Data Challenge

High Performance Computing

Very special thanks to Wolfgang Gentzsch and Burak Yenier for making the UberCloud HPC Experiment possible.

Simulation Platform Overview

NVIDIA AUTOMOTIVE. Driving Innovation

Welcome to the. Jülich Supercomputing Centre. D. Rohe and N. Attig Jülich Supercomputing Centre (JSC), Forschungszentrum Jülich

locuz.com HPC App Portal V2.0 DATASHEET

Supercomputing Resources in BSC, RES and PRACE

Mixed Precision Iterative Refinement Methods Energy Efficiency on Hybrid Hardware Platforms

Open Source CFD Solver - OpenFOAM

Deploying and managing a Visualization Onera

Parallel Computing with MATLAB

Technical bulletin: Remote visualisation with VirtualGL

CFD Applications using CFD++ Paul Batten & Vedat Akdag

10- High Performance Compu5ng

Remote & Collaborative Visualization. Texas Advanced Compu1ng Center

Formula One - The Engineering Race Evolution in virtual & physical testing over the last 15 years. Torbjörn Larsson Creo Dynamics AB

HPC-related R&D in 863 Program

Building an energy dashboard. Energy measurement and visualization in current HPC systems

High Performance. CAEA elearning Series. Jonathan G. Dudley, Ph.D. 06/09/ CAE Associates

Parallel Programming Survey

BC43: Virtualization and the Green Factor. Ed Harnish

extreme factory innovation at your fingertips

A-CLASS The rack-level supercomputer platform with hot-water cooling

Computational Fluid Dynamics (CFD) Markus Peer Rumpfkeil

Evoluzione dell Infrastruttura di Calcolo e Data Analytics per la ricerca

Appro Supercomputer Solutions Best Practices Appro 2012 Deployment Successes. Anthony Kenisky, VP of North America Sales

Bytes and BTUs: Holistic Approaches to Data Center Energy Efficiency. Steve Hammond NREL

How To Build A Supermicro Computer With A 32 Core Power Core (Powerpc) And A 32-Core (Powerpc) (Powerpowerpter) (I386) (Amd) (Microcore) (Supermicro) (

Extreme Computing: The Bull Way

HPC and Big Data. EPCC The University of Edinburgh. Adrian Jackson Technical Architect

FLOW-3D Performance Benchmark and Profiling. September 2012

ANSYS Computing Platform Support. June 2014

Jean-Pierre Panziera Teratec 2011

Agenda. HPC Software Stack. HPC Post-Processing Visualization. Case Study National Scientific Center. European HPC Benchmark Center Montpellier PSSC

HPC Cloud. Focus on your research. Floris Sluiter Project leader SARA

The Value of High-Performance Computing for Simulation

OMIQ srl. Advanced Post-Processing of Flow3D Calculations using Ensight. Federico Monterosso. Ingegneria di componenti e sistemi industriali

RWTH GPU Cluster. Sandra Wienke November Rechen- und Kommunikationszentrum (RZ) Fotos: Christian Iwainsky

HPC technology and future architecture

Design and Optimization of a Portable Lattice Boltzmann Code for Heterogeneous Architectures

PRIMERGY server-based High Performance Computing solutions

MIKE by DHI 2014 e sviluppi futuri

Fundamentals of CFD and Data Center Cooling Amir Radmehr, Ph.D. Innovative Research, Inc.

Energy Strategic Plan Los Angeles Community College District Community College League Conference

Transcription:

SysFera and ROMEO Make Large-Scale CFD Simulations Only 3 Clicks Away Benjamin Depardon SysFera Sydney Tekam Tech-Am ING Arnaud Renard ROMEO

Manufacturing with HPC 98% of products will be developed digitally in 2020 94 % of USA small and medium manufacturing companies have not yet adopted highperformance digital manufacturing Bill Magro Director of HPC software @ Intel

The problem ROMEO SMB HPC-Center

Making HPC accessible to SMBs One mission of the ROMEO HPC Center ROMEO is entitled to sell compute hours to SMBs Member of HPC-PME French initiative to help SMBs to start with HPC GENCI, INRIA, bpifrance, 10 tier-2 Regional HPC Centers (Equip@meso) But limited support/admin team, and no CFD engineer Industry needs Quality & service Simple access & usage Desktop environment

The solution SMB TECH-AM ING WEB PORTAL SYSFERA HPC-Center ROMEO

Outline Protagonists ROMEO SysFera TECH-AM ING Use Case: Simulating a mixing crystallizer Challenge Results Prospects

University of Reims Champagne-Ardenne (URCA) ROMEO Computing Center Multidisciplinary university arts, medicine, sport, more than 23 000 students a wide initial undergraduate studies program graduate studies and PhD program in link with research labs The ROMEO HPC Center is a platform hosted by URCA Member of the digital Research platform La Maison de la Simulation de Champagne-Ardenne Funded by European Community, French government, Champagne-Ardenne Council and Reims Metropolis high performance computing resources for both industrial and academic researchers in the region an in-depth expertise in different engineering fields: HPC, applied mathematics, physics, biophysics and chemistry. first Cuda Research Center in France European Centers National Centers equip@meso Integrated in the European HPC ecosystem link between large hardly accessible national centers and small research laboratories and SMEs of the region. member of the French Tier 2 network equip@meso managed by GENCI member of the European Platform ETP4HPC University / Regional Centers

SYSFERA Company created in 2010 Product Software Vendor SysFera-DS: Easy access to Cloud & HPC applications 11 persons, large majority of developers «Simplify access and administration of complex applications on hybrid clouds»

TECH-AM ING Company specialized in numerical simulation and complex CFD Has been working with Romeo since 2011 Combing its technical skills with ROMEO s considerable CPU power, TECH-AM ING performs complex systems modelling through HPC and CFD. Sydney TEKAM, CEO

Computational Fluid Dynamics (CFD) Numerical methods to predict fluid flow, heat transfer, mass transfer, chemical reactions Applications industrial processes (boilers, heat exchangers, combustion equipment, pumps, blowers, piping, etc.) Aerodynamics of ground vehicles, aircraft, missiles Film coating, thermoforming

Process for CFD simulations 1. Pre-processing Geometry Mesh Physical model 2. Simulation Iterative process Update parameters Quality control 3. Visualize & analyze

TECH-AM ING s Simulation Sugar crystallization process Customer: Process Mixing Solutions (PMS) Transform sugarcane or sugar beets syrup into sugar Classical process: heating the syrup during 4 hours Goals of the study Optimize the mixing process Heat transmission Viscosity consistency fan shape, force and power consumption Pumping / melting efficiency Syrup viscosity homogeneity Divide by 2 the heating duration (5 KW) Double the production (production peaks)

The Crystallizer Ø5m x 5m high tank (16ft) 100 cubic meters (26 Kgal) Vertical heat exchangers 227 tubes one meter-high tubes Moving fans 20 million cells mesh minimum if we want to model precisely the tubes Simulated process estimated duration: 2 hours CFD simulation using ANSYS Fluent (thanks) Process details and system geometry courtesy of PMS (http://processmixingsolutions.com)

How to make it 3 clicks away? SMB TECH-AM ING WEB PORTAL SYSFERA HPC-Center ROMEO

ROMEO HPC Tesla Cluster Compute Display 5 th 3131 MFLOPS/W Bull Cool Cabinet Door 151 th 254.9 Tflops Linpack 260 NVIDIA Tesla K20X accelerators 130 Bull servers bullx R421 E3 Bull AE & MPI 260 INTEL Ivy Bridge E5-2650 v2 Processor (2,66 GHZ / 8c), non-blocking Mellanox Infiniband, Slurm, 88 To Lustre, 57 To home, 100 To Storage, 42 Westmere nodes Installation et classements automne 2013 Big Data, on-demand and remote Quadro 5800, 6000, K6000, K2 Specific solutions VirtualGL NVIDIA GRID + Citrix Virtualisation Windows Scalable Graphics Compatible / standard solution X VNC Batch Scheduler integrated Multi-nodes rendering

CFD as a Service Pre-Processing Run Post-Processing View App List of options Option one Option two Input View Submit

CFD as a Service Through SysFera-DS web portal Batch tests & simulations Post-processing & visualization

The run Parameters Turbulence (k-eps model) Heat exchange physics Moving meshes For 4 minutes of simulation 24 hours of processing 1024 cores Whole process would take 1 month of processing

Demo (video) https://romeo.univ-reims.fr/news/188 Journée

Benefits and impacts TECH-AM ING & other SMBs Simple and immediate access to CFD tools New users are autonomous Nothing to install on the user side ROMEO Less support needed Easy addition of new & tailor made services Attract more SMBs Other applications for SMBs Sysfera-DS available for academic users

Sydney TEKAM President of TECH-AM Ingenierie s.tekam@techam-ingenierie.com http://techam-ingenierie.com Benjamin DEPARDON VP Business Development benjamin.depardon@sysfera.com http://www.sysfera.com Arnaud RENARD Chef de Projet ROMEO HPC center arnaud.renard@univ-reims.fr https://romeo.univ-reims.fr