Slides for the Presentation by Dr. Mary K

righteousgaggleΔιαχείριση Δεδομένων

31 Ιαν 2013 (πριν από 4 χρόνια και 7 μήνες)

115 εμφανίσεις


SRNL Computational Sciences


Dr. Mary K Harris

SRNL
-
STI
-
2012
-
00199

2









SRNL Computational Sciences

Applied Computational

Engineering & Statistics

(Patricia Lee, Ph.D)

Computational Sciences

Mary K Harris, Ph.D
-

Director

Chief Information Officer

Mary K Harris, Ph.D

Computing Program

Manager & CISSP

(Cynthia Holding
-
Smith)

Computing Program

Manager

(John Longo)

Administrative Assistant

(Carolyn Ervin)

Computational Engineering

& Sciences

Steve Hensel, Ph.D

Scientific Computing

Phil Moore, Ph.D

Process Modeling &

Computational Chemistry

(Steve Hensel, Ph.D)

Software Engineering

(Ben Torkian, Lead)

High Performance

Computing/Linux Systems

Engineering

(Tommy Ansley, PM)

3









SRNL Computational Technical Staff

Approximately 75 Computational Professionals

Engineers


chemical


mechanical


nuclear


environmental


civil

Computer Scientists

Meteorologists

Statisticians

Chemists

Experience spans broad spectrum of applications
supporting Environmental Stewardship, National Security,
and Energy Security


4











Applied Statistics

Statistical sampling for process qualification and uncertainty determination

Material control & accountability and associated uncertainties

Monte Carlo simulations and statistical forensics

Computational Engineering

Computational fluid dynamics and heat transfer

Structural and coupled multi
-
physics analysis

Radiation transport modeling

Environmental Modeling


Groundwater, geochemical, risk assessment, and regulatory

Chemical Process Modeling


Nuclear materials processing


Flow sheet and systems modeling

Atmospheric Modeling


Local, regional and international, CO2 risk assessment, climate, emergency response

Computational Chemistry


Chemical Reactions, Separations, and Molecular/Compound Design


NMR and XRD Structure Validation


Structural Stability and Kinetic Processes









Key Technical Competencies

5









Key Resources: Hardware, Software


Hardware

HPC (managed by Scientific Computing)

High end Linux workstations (managed by Scientific Computing)

High end Windows workstations (managed by IT)


Software

SAS, JMP (statistics)

Fluent,
Abaqus
,
Comsol
,
Patran
/Thermal, MCNP (
eng.

modeling)

Gaussian,
Castep
,
Dmol
,
Cosmotherm
, Wien, ADF (Comp.
Chem
)

ACM, Aspen Plus, OLI, Verse, Extend (Process Modeling)

other not managed by CES,
Porflow
,
Goldsim

6









Key Products/Services/Functions


We solve problems


Product is typically a report...rarely a
model to be used by the client

Broad Range of Modeling/Simulation and Calculations using
appropriate rigor and computational tools

from atoms to
flowsheets
/systems

from spreadsheet to sophisticated commercial
software

from windows PC’s to multicore Linux compute
servers


Unclassified and Classified work

Strategic partnerships with customer base

7









Key Technical Competencies

High Performance Computing (HPC)


Computer modeling, Visualization, Distributed computations

Scientific Applications Management and Development

Laboratory Information Management Systems (LIMS), Open Source (Python, R), Commercial (Matlab, LabView,
Fluent, COMSOL, PORFLOW, PDM
-
Link), Python
-
Java
-
Fortran
-
C software development, Graphical User
Interfaces, Integration of Windows Scientific Applications onto SRS Desktop

Archival, Retrieval and Analysis of Scientific Datasets

Remote data collection (Weather data, Insitu, Sodar), TADAAS Data Archive and Analysis (DHS/DNDO),
Enterprise archive and version control of SRNL software and documents

Networks for Research and Academic Collaboration

Establish DHS network in SRS, Open collaboration network (srnl.org), Cyber Security policies, procedures, and
applications, DHS
-
NNSA cyber systems and applications support

SRNL Web Development

RAMPAC (rampac.energy.gov), Hydrogen Storage Engineering Center of Excellence (hsecoe.srs.gov)

PC Controllers and SRNL Windows Applications

Footprints helpdesk ticket system, TIVOLI backups, PC Desktop support

Cyber Security


Policy, Risk Assessments, Scanning, Patching







8









HPC Facilities

SRNL Linux Cluster
-

1210 cores, 35 TB memory, 200 TB disk storage

DOE Unclassified, Disk storage for Linux and Windows Desktops

Atmospheric Technologies Group compute cluster

Linux Scientific Workstations


DHS TADAAS HPC Facilities (Prototype in Progress)

Linux
Cluster
-

114 cores, 19 TB memory, 16 TB disk storage

Provide authenticated remote access from multiple external sites to
centralized SRNL
-
DNDO data archive and analysis system

Meet DHS cyber security requirements


HPC Open Computing Facilities (March 2012)

Linux Cluster, Open Source applications, srnl.org domain, 10 Gigabit network

Authenticated Open Access for collaboration with universities, laboratories
and industry




Key Facilities

9









Linux HPC systems
(hardware, software and high
-
speed
network)

Scientific applications
(PorFlow, Fluent, Matlab, COMSOL)

LIMS
(Labvantage, Labtronics, Oracle)

ATG WINDS Data Collection
(Python, Postgres)

Visualization
(DOE ASCEM
-
Vis, VisIt, MatPlotLib)

Linux, Windows, IPad and Web Software Development
(Weather Applications, RAMPAC, Cercla, MPF)

Open
-
source software for reducing costs, removing licensing
barriers and expanding computing capability

(Scientific Python,
R, OpenFOAM, Sierra, Delft3D)


Key Services

10









Example
-

ATG Weather System Data Flow

11









Weather Application Conversion Project

Solaris and Windows

Oracle

C

PL/SQL

Fortran

TCL

TK

Oracle Forms

IDL

Linux

PostgreSQL

C

pg8000,
pygresql

Fortran

Python

Tkinter

wxPython

Matplotlib


12









ATG Legacy User Interface

13









ATG Weather System Web Demonstration

14









Key Initiatives

High Speed 10 Gigabit Connection from SRNL

to the National Lambda Rail and Internet2


Collaboration User Facility in Performance
-
Optimized Datacenter (POD)

Integrate Graphical Processing Units (GPUs) into SRNL Computing
environment

Collaborate with other national laboratories and universities by establishing
shared computing facilities

Integrate Open Source Software into the SRNL Scientific
C
omputing
environment

Increase applications and software support for SRNL departments

Balance staffing and expertise in support of SRNL Scientific Computing
need
s

More opportunities for Computer Science and Engineering internships

15









A Collaborative Path Forward for SRNL Scientific Computing

Provide resources for collaborating with academic institutions
in teaching Clean Energy concepts and technologies using
computer models and simulations

Actively
participant in
DOE
ESNet

and
Internet2 activities in
association with other National Laboratories & Academic
institutions

Facilitate dialog through video conferencing







16









Areas SRNL Scientific Computing is projected to increase


Cyber security


unclassified and classified


Technical Network Engineering Support


Software applications


Database management


Multi
-
level support for collaborative computing
environments


Linux administration



17









Contact Information

Dr. Mary Harris

mary.harris@srnl.doe.gov


Dr. Steve Hensel

steve.hensel@srnl.doe.gov


Dr. Phil Moore

phil.moore@srnl.doe.gov


Supercomputing 2011


Seattle, WA

18









Your future looks prosperous


Employment


growing faster than the average for
all occupations

Increase 48.78% 2008
-
2018


Broad range of formal education levels


Part of everyday life


You will be in high demand


Telecommuting will become routine


Salaries competitive and growing

Education level and area of expertise



19









Who employs Computer professionals?


Very wide range of industries

Telecommunications, financial, insurance, business
management, power, schools, government, national labs, film
industry, self
-
employed, manufacturing industry, and more


http://www.bls.gov/oco/ocos305.htm

Bureau of Labor Statistics Computer Network, Systems, and
Database Administrators