gleb radchenko , anastasiya shamakina south ural state university (susu)

29
A Service-Oriented Approach of Integration of Computer-Aided Engineering Systems in Distributed Computing Environments Gleb Radchenko, Anastasiya Shamakina South Ural State University (SUSU) Supercomputer Simulation Laboratory (SSL) 30 May, 2012, UNICORE Summit, Dresden.

Upload: judith

Post on 25-Feb-2016

44 views

Category:

Documents


0 download

DESCRIPTION

A Service-Oriented Approach of Integration of Computer-Aided Engineering Systems in Distributed Computing Environments. Gleb Radchenko , Anastasiya Shamakina South Ural State University (SUSU) Supercomputer Simulation Laboratory (SSL) 30 May, 2012, UNICORE Summit, Dresden. - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

A Service-Oriented Approach of Integration of Computer-Aided Engineering Systems in

Distributed Computing Environments

Gleb Radchenko, Anastasiya ShamakinaSouth Ural State University (SUSU)Supercomputer Simulation Laboratory (SSL)

30 May, 2012, UNICORE Summit, Dresden.

Page 2: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

South Ural State University

2

Page 3: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

Communication networks:

System network: 3D torus, 60 Gbit/s InfiniBand QDR, 40 Gbit/sGigabit Ethernet 3

SKIF-Aurora SUSU Supercomputer Specifications

Ranking TOP50 CIS list: 3rd Ranking TOP500: 87 (June 2011)117 Teraflops

8832 computing cores

RAM: 9 TByte

Disk space: Intel SSD 108 TByte

Page 4: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

4

Moscow

ChelyabinskYekaterinburg

St .Petersburg

TomskNovosibirsk

Nizhniy Novgorod

Page 5: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

Distributed Problem-Oriented Environments

5

Page 6: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

Goal:The principal objective is to develop a

technology allowing to take into account the specifics of the problem-oriented subject areas while providing the recourses of distributed computing environments.

This technology aimed to create “intelligent” middleware providing users with easy, transparent and secure access to distributed computing resources and allowing them to solve specific classes of applied problems. 6

Distributed Problem-Oriented Environments

Page 7: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

A Problem-Oriented Environment

7

Model

Application programmer

Grid

UserDevelopment

Define Parameters of

Simulation

Get Results of Simulation

Run the Simulation

Page 8: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

The Basic Definitions: Job

8

Creation or correction of a

geometric model

Creation or correction of a mesh

Creation or correction of problem

definition

Problem Simulation in CFX-Solver

Visualization in CFX-Post

Values of the optimization

criteria are not satisfactory

The accuracy of the calculation is not satisfactory

Job defines the process of model simulation

Job is a set of Tasks organized as a workflow aimed at achieving of a useful result.

Page 9: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

The Basic Definitions: Tasks

9

DesignModelerService

• Task represents a solution for some part of a job

• Task describes the process of transformation of input parameters into output parameters

Con

vert

er

Problem-oriented parameters

Template

Converter

Geometry

Input parameters Transformation Output

parameters

Script

Page 10: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

“Simulating Flow in a Static Mixer” Job and Tasks

example

10

Design Modeler

CFX-MeshCFX-Pre

CFX-Solver

CFX-Post

Creation or correction of a

geometric model

Creation or correction of a mesh

Creation or correction of problem

definition

Problem Simulation in CFX-Solver

Visualization in CFX-Post

Values of the optimization

criteria are not satisfactory

The accuracy of the calculation is not satisfactory

Page 11: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

The Basic Definitions: Resources

11

CFX-Solver

Hardware resources

Software resources

Licenses resources

Target system

Resource is a hardware and software required to perform a Task

Page 12: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

Distributed Virtual Test Bed

12

Page 13: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

13

Distributed Virtual Test Bed

ResourceBroker

UNICORE

DiVTB Web Interface

A Driver

Simulation Results

EngineerDistributed Virtual Test Bed(DiVTB ) includes an interface for a CAE-problem

statement; a driver (a set of software tools

enabling the use of grid resources for virtual experiment);

a set of grid services (a set of supercomputers in a distributed computing environment, with the installed software components)

Page 14: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

14

Distributed Virtual Test Bed

Features:Providing a safe and transparent web-access to

the resources of a distributed computing environment

User do not need to know about the distributed nature of the computing environment

User do not need desktop application to perform simulation

Page 15: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

CAEBeans System

15

Page 16: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

CAEBeans Constructor

16

Page 17: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

CAEBeans Portal

17

Page 18: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

CAEBeans Server

18

Page 19: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

CAEBeans Broker

19

Page 20: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

CAEBeans ResourcesCAE-Resources – UNCORE/X Site with several

special applications

20

CAE-system

UNICORE

Implementation

Target System Interface

Interface

UN

ICR

E

App

licat

ion

Page 21: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

Example: LS-Dyna CAE-Resource

21

<!-- LS-Dyna application --> <idb:Application> <jsdl:ApplicationName>LS-Dyna</jsdl:ApplicationName> <jsdl:ApplicationVersion>1.0</jsdl:ApplicationVersion> <idb:ScriptTemplate> <idb:Description> Template for LS-Dyna Simulation Invocation. Fields: SOURCE source file name DOUBLE_PRECISION double precision or not (y/n) (*) PROCESSORS nomber of cores (*) MEMORY total memory (*) MEMORY2 memory per core (*) (*) - optional field, default value is provided. </idb:Description> <idb:Invocation name=""> <idb:Body>![CDATA["ls-dyna.sh" +S<SOURCE> +DP<DOUBLE_PRECISION> +NP<PROCESSORS> +M<MEMORY> M2<MEMORY2>]]</idb:Body> </idb:Invocation> ... </idb:ScriptTemplate> </idb:Application>

Page 22: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

DiVTB for practical engineering problems solution

22

Page 23: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

DiVTB for common CAE-systems

We developed DiVTB to solve problems by using common CAE systems

ANSYS CFX,

ANSYS Mechanical,

ABAQUS,

DEFORM,

LS-DYNA.

23

Page 24: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

Vortex Flow Meter DiVTB

24

Pressure distribution in flow tube of vortex flow meter

Page 25: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

DiVTB «Simulation of Tube Ovalization during Hardening»

25

Customer:JSC Chelyabinsk Tube-Rolling Plant

Purpose of study:simulation of tube deformation during hardening during production on tube rolling plants

Result:rejection rate reduced

by 10%

Temperature distribution during water sprayer pipe cooling

Page 26: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

Simulation of Deformation of Flat Knitted Products Structure

26

The mechanism of “dressing" jerseys on the

dummy

Customer:Kyshtym knitwear Ltd.

Purpose of study:development of new methods of clothes design that take into account properties of the tissue

Result:automation of design process time to design new models

reduced product range growthvolume of output increased by

20%

Equivalent Mises stress (Pa)

Page 27: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

Deformation and Destruction of Fabric Bulletproof Vest by a Bullet Strike

The dynamic interaction of ballistic fabric with a bullet

Tissue concealable bulletproof vest

27

Customer: UAB “FORT Technology”

Purpose of study:development of new body armor design

Result:• profit margins increased from 10%

to 30% due to reduction of field experiments and use of virtual prototyping

• bulletproof vests are on arms of special units of the Russian Federation and foreign countries (Germany, France, Israel, Poland, Czech Republic, etc. )

Personal Safety Issues

Page 28: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

What’s next?Development of billing services for DiVTBDevelopment of CAEBeans Constructor Web-

applicationDevelopment of scheduling algorithms for

CAEBeans BrokerDevelopment of remote 3D-visualization servicesIntegration of authorization and authentication

with UNICORE servicesIntegration with UNICORE workflow systemIntegration with multi-criteria optimization

systems

28

Page 29: Gleb Radchenko ,  Anastasiya Shamakina South Ural State  University (SUSU)

Questions?

29

http://supercomputer.susu.ac.ru/en/