Towards Multiscale Computing Tools based on GridSpace

Towards Multiscale Computing Tools based on GridSpace

Katarzyna Rycerz, Eryk Ciepiela, Daniel Harężlak, Marian Bubak

ACC Cyfronet and Institute of Computer Science, AGH, Krakow, Poland

dice.cyfronet.pl

Work supported by MAPPER: Multiscale Applications on European e-Infrastructures, http://www.mapper-project.eu , „e-Infrastructures”

Project Director: Alfons Hoekstra, Amsterdam University

Cracow Grid Workshop 2010

Overview

• Multiscale simulations - overview• MAPPER motivation and

architecture• GridSpace – short reminder from

yesterday• Preliminary experiment with

multiscale application in GridSpace• Demo of the experiment

Multiscale SimulationsConsists of modules of

different scale

Examples – e.g. modelling:virtual physiological

human initiative

reacting gas flows

capillary growth

colloidal dynamics

stellar systems

and many more ...

virtual physiological humanvirtual physiological human fusionfusion hydrologyhydrology

nano material sciencenano material science computational biologycomputational biology

the reoccurrence of stenosis, a narrowing of a

blood vessel, leading to restricted blood flow

MAPPER architectureDevelop computational

strategies, software and services

for distributed multiscale simulations across disciplines

exploiting existing and evolving European e-infrastructure

Deploy a computational science infrastructure

Deliver high quality componentsaiming at large-scale, heterogeneous, high performance multi-disciplinary multiscale computing.

Advance state-of-the-art in high performance computing on e-infrastructures

enable distributed execution of multiscale models across e-Infrastructures,

• Easy access using Web browser• Experiment workbench

• Constructing experiment plans from code snippets

• Interactively run experiments• Experiment Execution Environment

• Multiple interpreters• Access to libraries, programs and

services (gems)• Access to computing infrastructure:

Cluster, grid, cloud• Example applications using GS

• Binding sites in proteins• Analysis of water solutions of

aminoacids• Experience

• Virolab project• PL-Grid NGI

GridSpace

Preliminary experiment with multiscale application in GridSpace

• Multiscale dense stellar system simulations (from MUSE; http://www.muse.li)

• Two modules with different scales:– stellar evolution (macroscale)– stellar dynamics - N-body

simulation (mesoscale) • Data management

– masses of evolving stars sent from evolution (macroscale) to dynamics (mesoscale)

– no data is transmitted from dynamics to evolution

– dynamics should not outpace evolution

21

2

1

data

data

evolution dynamics

MUSE application

1000 ...

10021001

2000 ...

simulation time

1

2

No. of steps

Interactions between components in our experiment • We use a special communication bus

(called HLA) to synchronize simulation modules with time management

• Time management – Simulation modules called federates– regulating federate (evolution)

regulates the progress of the constrained federate (dynamics)

• federates exchange data with time stamps

• The furthest point in time which the constrained federate can reach at a given moment (LBTS) is calculated dynamically, according to the position of the regulating federate on the time axis

LBTS - Other federates will not send messages before this time.

Federate may only advance time within this interval

Federate’s current logical time.

Federate’s effective logical time.

Federate may not publish messages within this interval

Federate’s current logical time.

t=0

Lookahead

Constrained federate(dynamics)

Regulating federate (evolution)

Dynamics Evolution

HLA communication bus

Wrapping simulation models as software components

To enable users to steer the behavior of the simulation from outside we wrap simulation models into software components

We use the H2O framework• simulation modules can expose remotely accessible external

interfaces• implementations of simulation models are wrapped and placed

inside pluglets• containers for pluglets are called kernels• pluglets are deployed into kernels

Remote node

H2O kernelH2O pluglet

implementation of simulation

model

Start/stop

Change time policy

Switch on/off data exchange

Demo experiment

H2O kernel

node A

H2O kernel

node B

Ruby script(snippet 1)

Run PBS job

allocate nodes

start H2O kernelsGridSpace

user

PBS run job (start H2O kernel)

Ruby script(snippet 1)

Demo experiment

H2O kernel

node A

H2O kernel

node B

Jruby script(snippet 2)

Asks selected components to join simulation system

Asks selected components to publish or subscribe to data objects (stars)

Asks components to set their time policy

Determines where output/error streams should go

HLA communication

join federation

join federationsubscribe

publishbe constrained be regulating

Dynamics HLAComponent

EvolutionHLAComponent

GridSpace

set streamingset streaming

user

create components

Ruby script(snippet 1)

GridSpace

H2O kernel

node A

H2O kernel

node B

Asks components to start

Alters the time policy at runtime

Stop

Dynamics HLAComponent

EvolutionHLAComponent

HLA communication

startstart

unset regulation

Star data objectStar data object

Demo experiment

unset constrained

Jruby script(snippet 2)

user

Jruby script(snippet 3) Jruby script

(snippet 4)

stop

stop

Dynamics

view

Evolution

view

Out/err

Ruby script(snippet 1)

Demo experiment

H2O kernel

node A

H2O kernel

node B

Ruby script(snippet 5)

Delete job

stop H2O kernels

release nodesGridSpace

user

PBS Delete job ( stop H2O kernels)

Ruby script(snippet 1) Ruby script(snippet 1) Ruby script(snippet 5)