DTIC ADA512448: The Resource. Spring 2003

33 Pages·2.2 MB·English

by  Defense Technical Information Center

Report Documentation Page Form Approved
OMB No. 0704-0188
Public reporting burden for the collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and
maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information,
including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington
VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to a penalty for failing to comply with a collection of information if it

does not display a currently valid OMB control number.
1. REPORT DATE 3. DATES COVERED
2003 2. REPORT TYPE 00-00-2003 to 00-00-2003
4. TITLE AND SUBTITLE 5a. CONTRACT NUMBER
The Resource. Spring 2003
5b. GRANT NUMBER
5c. PROGRAM ELEMENT NUMBER
6. AUTHOR(S) 5d. PROJECT NUMBER
5e. TASK NUMBER
5f. WORK UNIT NUMBER
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGANIZATION
U.S. Army Engineer Research and Development Center,ATTN: ERDC REPORT NUMBER
MSRC HPC Resource Center,3909 Halls Ferry
Road,Vicksburg,MS,39180-6199
9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSOR/MONITOR’S ACRONYM(S)
11. SPONSOR/MONITOR’S REPORT
NUMBER(S)
12. DISTRIBUTION/AVAILABILITY STATEMENT
Approved for public release; distribution unlimited
13. SUPPLEMENTARY NOTES
14. ABSTRACT

15. SUBJECT TERMS
16. SECURITY CLASSIFICATION OF: 17. LIMITATION OF 18. NUMBER 19a. NAME OF
ABSTRACT OF PAGES RESPONSIBLE PERSON
a. REPORT b. ABSTRACT c. THIS PAGE Same as 32
unclassified unclassified unclassified Report (SAR)
Standard Form 298 (Rev. 8-98)
Prescribed by ANSI Std Z39-18
From the Director's chair
While this is only my second newsletteras Directorofthe Engineer Research and
DevelopmentCenterMajorShared ResourceCenter(ERDCMSRC), it is my lD,h
year in theCenter, which means I've been hereas long as it has. In all that time
there hasbeen only one fundamental shift in the way weget the business ofthe
MSRCdone -theshift in the mid-90s from aprimarilyGovernment-owned,
Government-operatedorganizationtoan organizlIlion slaffed largely by an
integration conlractorled by theGovernment.
This wasa big changeaimed at bringing tlexibility into theway theCenters wereoperated. The result was the four
Centers wehave now,each a major force in theprovision ofcomputational e,xpertiseand resourcesenabling the
DcpanmcllI ofDefense(000) wurfighter mission.
That approach served us- and you, ourusers- very well. However, the passageoftime, change in requirements,
and continual evolutionoftheGovernment-colllractor landscape brought us tothe point thm we needelltoconsider
..ll1other fundamental shift in the way we get the business oftheMSRCdone.
On April 1,2003.that shift wascompleted.TheERDCMSRCcompletely restructured its servicecontracts. moving
from the largeumbrellaofasingle integration contract toaconstellationofcontracts that provides clusters of
expertiseand functionality for theCenter.Thethree majorareasofthesecontracts are servicedelivery (system
administration. technology planning,etc.). serviceSUPPOl1 (helpdesk. computational science. scientificvisualiza
tion, etc.),and maintenance.
Creating a service contract structure like this affords us avariety ofcritical abilities as wecontinue to striveto
improveourservice model. Most importantly, it will allow the ERDCMSRCtorespond more l1exibly 10thechanging
demandsofourusercommunity. resultingultimately in a higher level ofservicefor ourusercommunily.
Forus. this is asignificantchange that will beaccompilnied by the usual growing pains as we struggle to reshape
how we getlhingsdone. Foryou, however, the mechanics ofthischange will be virtually invisible. It is ourgoal
that lheonly effectsyou see are more services provided in the way you want them, when you want them.
A'? we movethrough thisperiodoftmnsitiOn within the Center, I encourageyou [0continuetogive us feedback
about how wecan helpyou moreeffectively.The UsersAdvocacy Group is avery effective way togelyour input to
us; wewelcomeandencour..lgeyou to respond individually to usas well.Theonly way totruly match whal we
provide10whatyOll need isopen,honest,cOJllinuouscommunication.
John West
Director,ERDCMSRC
About the Cover;
Seismic signature simulations for unattended ground sensor systems and numerical simulation of ground vehicle
tracking systems using seismic signatures (see article, page 6). Cover design by the EAOC MSRC Scientific
Visualization Center.
Features
An HPC-Enabled Virtual Proving Ground for Seismic Unattended
Ground Sensor Networks................................................................... 6
SC2002 “From Terabytes to Insights” ....................................................... 9
PET Highlights....................................................................................... 12
Building on Previous Strategies to Create a Synthetic Application
of Benchmarking .............................................................................. 14
Scientific Visualization Center Technology Update ............................... 17
Army Science Conference – “Transformational Science &
Technology for the Army...a race for speed and precision” .............. 18
Departments
announcements......................................................................................................................2
upcoming events ....................................................................................................................4
off-campus..............................................................................................................................5
technology update................................................................................................................20
Technology Enhancements in the ERDC MSRC Computational Environment ....................20
ERDC MSRC Prepares to Assist Users When Cray X1 Arrives ...........................................21
community outreach.............................................................................................................22
Job Shadowing in Computer-Related Fields ....................................................................22
ERDC MSRC Staffers Share Career Knowledge with Students...........................................23
visitors..................................................................................................................................25
acronyms..............................................................................................................................28
training schedule..................................................................................................................28
p.12
p.6
p.9 pp..1188
The Resource, Spring 2003 ERDC MSRC 1
announcements
○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
“Army Newswatch” to Feature DoD Supercomputer Center at ERDC
By Rose J. Dykes
“Army Newswatch,” SRTV’s
premier television news pro-
gram, showcases the Army as
it conducts its many roles and
missions in support of the
Nation. It takes a compre-
hensive look at what is hap-
pening throughout the Army
and focuses on the issues, the
equipment, and the people
that make the Army what it
is today. This award-winning
television newscast is a
biweekly production.
Soldiers Radio and Television reporter Hank Heusinkveld films
John E. West in the ERDC Supercomputer Center.
Soldiers Radio and Television (SRTV) reporter
Hank Heusinkveld visited the MSRC on
October 24, 2002, and interviewed John E.
West, Director, while walking through the
Joint Computing Facility. The news feature
on the MSRC will eventually air at the Penta-
gon, at worldwide Army installations, and on
several hundred cable systems in the United
States.
Other ERDC news to be featured along with
that of the Supercomputer Center include
ERDC as the 2002 Army Research and Devel-
opment Organization of the Year, force
protection and antiterrorism research and
development, the TeleEngineering Operations
(Left to right) Wayne Stroupe, ERDC Public Affairs office, Hank
Center, and the new Ship-Tow Simulator.
Heusinkveld, SRTV, and John E. West, ERDC MSRC Director,
visit before filming the MSRC news feature.
The ERDC MSRC welcomes comments and suggestions regarding The Resource and invites
article submissions. Please send submissions to the following e-mail address:
[email protected]
2 ERDC MSRC The Resource, Spring 2003
announcements
○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
Users Advocacy Group – A Forum for DoD HPCMP Resource Users
By Rose J. Dykes
Dr. Stacy Howington, ERDC Coastal and Hydraulics
Laboratory, is the newest member of the high- Dr. Stacy Howington
received his B.S. and
performance computing (HPC) Users Advisory Group
M.S. degrees from
(UAG). This group was formerly known as the Shared
Mississippi State
Resource Center Advisory Panel or SRCAP. The UAG
University in 1983 and
mission has recently been changed as follows:
1988, respectively, in
! Provides a forum for users of the DoD High Perfor- civil engineering. In
mance Computing Modernization Program’s 1997, he received his
(HPCMP) resources to influence policies and Ph.D. in civil engineer-
practices of the Program. ing from the University
! Facilitates the exchange of information between of Colorado at Boulder.
the user community and the HPCMP.
! Serves as an advocacy group for all HPCMP users. Dr. Howington works in the ERDC Coastal and
! Advises the HPC Program Office on policy and Hydraulics Laboratory where he does modeling
operational matters related to the HPCMP. of fluid flow and constituent transport in
groundwater and surface water systems. He has
The Army, Navy, and Air Force each appoint four
been associated with the Environmental Quality
people to represent them as their service members; one
Modeling and Simulation and Climate/Weather/
additional member is selected to represent other DoD
Ocean Modeling and Simulation Computational
agencies. The services are encouraged to appoint
Technology Areas for several years.
members who are active users in the HPCMP to best
represent the user community. The members serve
2-year renewable terms.
Meetings are scheduled at least twice a year. Representatives from the Shared Resource Centers, although not
members, have a standing invitation to attend the meetings. A user should contact one of the four representatives in
his same service. If a user does not work for the Army, Navy, or Air Force, he should contact the representative
from other DoD agencies. Users can always contact the Program Office to get the names of their service representa-
tives. Appropriate issues for users to take to UAG members are ones that need to be brought before the entire
Program for the good of the whole user community – not things such as a machine is not working as it should at a
particular site.
The full list of HPC UAG members is found below.
Air Force Navy
! Jerry Boatz (S&T), [email protected], ! Joe Gorski (S&T), [email protected], (301)
[phone] 227-1930
! Bonnie Heikkinen (T&E), ! Ed Neal (T&E), [email protected], (301)
[email protected], [phone] 757-1781
! John Martel (T&E), [email protected], ! Jeanie Osburn (S&T), [email protected], (202)
[phone], Extension 3368 767-3885
! Stephen Scherr (S&T), ! Alan Wallcraft (S&T),
[email protected], [phone] [email protected], [phone]
Army Other
! Stacy Howington (S&T), ! Steve Finn (DTRA), [email protected], (310)
[email protected], [phone]
634-2939
! Michael J. Reil (T&E), [email protected], (410) NOTE: S&T (Science and Technology)
278-9474 T&E (Test and Evaluation)
! Stephen Schraml (S&T), [email protected], DTRA (Defense Threat Reduction Agency)
[phone]
! Jackie Steele (T&E), [email protected],
[phone]
The Resource, Spring 2003 ERDC MSRC 3
announcements
○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
Two CSC Employees at ERDC MSRC Honored for HPC Work
By Ginny Miller
Two Computer Sciences Corporation (CSC) employees were
recognized by CSC for their contributions to the company’s high-
performance computing (HPC) efforts at the ERDC MSRC. Robert
Scudamore, Vice President of CSC’s HPC Center of Excellence,
presented the awards November 18, 2002, during the 15th annual
Supercomputing Conference (SC2002) in Baltimore, Maryland.
Carrie Mahood received CSC’s High Performance Computing
Outstanding New Employee Award. The award was in recogni-
tion of Mahood’s immediate contributions to the Computational
Science and Engineering (CS&E) group at the ERDC MSRC. Her
work included a lead role involving testing and analysis on a
project with the U.S. Army Corps of Engineers New Orleans
District, as well as conducting multilevel parallel programming workshops at the Arctic Region Supercomputing Center
in Fairbanks, Alaska, and SC2002.
Mahood joined CSC in October 2001 as a computational scientist at the ERDC MSRC. She is a 1999 graduate of East
Texas Baptist University in Marshall, Texas, where she earned a bachelor’s degree in mathematics and computer
science. Mahood graduated in August 2001 from Texas Tech University with a master of science in mathematics.
Scudamore also presented a Technical Contribution Award to
Robert Alter in recognition of his continued contributions to the
CS&E group as the resident expert of Message Passing Interface-
Input/Output, for which he provided invaluable support to the
Seismic Wave Propagation in Parallel Topography code.
Alter received a bachelor of science in mathematics from Boise
State University in Idaho in 1977. He also received a bachelor of
science in geophysics from Boise State University in 1983. A
former high school mathematics teacher, Alter has worked as an
exploration geophysicist for Amoco Oil Production Company and
spent 14 years as an oceanographer at the Naval Oceanographic
Office at Stennis Space Center. He joined CSC in December 2000.
upcoming events
○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
2003 Users Group Conference – June 9-13, 2003, DoubleTree Hotel, Bellevue,
Washington, hosted by the DoD High Performance Computing Modernization Program.
SC2003 Conference – “Igniting Innovation,”
November 15-21, 2003, Phoenix, Arizona.
4 ERDC MSRC The Resource, Spring 2003
off-campus
○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○
MSRC Team Member Presents Seminar at USM
By Rose J. Dykes
Dr. Deborah Dent, Deputy Director of the Information Technology
Dr. Fred Tracy presents a seminar at USM.
Laboratory (ITL), ERDC, and Dr. Fred Tracy, MSRC Team, ERDC,
went to the campus of the University of Southern Mississippi
(USM), Hattiesburg, Mississippi, on November 1, 2002, where
Dr. Tracy presented a seminar entitled “Multi-Level Parallelism
(MLP) – An Alternative Parallel Paradigm.”
MLP is a new, lightweight approach to expressing the semantics of
a parallel program where communication among processors is
done through shared variables as in OpenMP rather than sends,
receives, broadcasts, reductions, etc., as in Message Passing
Interface.
The seminar was part of the class SC 740 Graduate Seminar taken
by graduate students in the Program of Scientific Computing and related fields of study at USM. Approximately
20 students and faculty attended.
After the seminar, Drs. Dent and Tracy visited with faculty members and discussed plans for possibly working on joint
projects and future visits to ITL. The Coordinator of the Program of Scientific Computing at USM is Dr. Joseph Kolibal.
IEEE Visualization 2002
By Paul Adams
Three team members of the ERDC MSRC Scientific Visualiza-
tion Center, Paul Adams, Dr. Michael Stephens, and
Richard Walters, attended IEEE Visualization 2002 on
October 27 – November 1, 2002, in Boston, Massachusetts.
Dr. Stephen Wolfram, a well-known scientist who received
his Ph.D. in theoretical physics in 1979 at the age of 20,
was the keynote speaker.
In 1986 Dr. Wolfram created Mathematica, a technical com- (Left to right) ERDC MSRC SVC Lead Paul Adams and
puting tool for the scientific research community that is now Dr. Michael Stephens and Richard Walters, SVC team
used worldwide. A New Kind of Science is his latest book. members, attended IEEE Visualization 2002.
In this book and in his keynote speech at the conference,
Dr. Wolfram talked about cellular automata, which produce shaded images on grid patterns according to certain rules.
He showed that incredible complexity can arise from simple systems and rules.
Other topics of interest at the conference included the following:
! Commodity-Based Cluster Visualization – How to manage and use scalable display walls.
! High-Quality Volume Graphics on Consumer PC Hardware – Using consumer graphics card for volume rendering.
! State of the Art in Data Representation for Visualization – Using signal processing to take laser-scanned data and
reduce it to a manageable size.
! Out-of-Core Algorithms for Scientific Visualization – How to handle data sets that are larger than main memory.
! Interactive Rendering of Large Volume Data Sets – Using wavelet compression and Level-of-Detail to view data sets
too large to be stored on a PC.
! Exploring Scalar Fields Using Critical Isovalues – Critical isovalues are those that are a minimum, saddle, or
maximum. By creating a critical isovalue locating program, one can ensure not missing any important features.
The listing of the critical isovalues can also be used to create transfer functions for volume rendering. However,
with too many isovalues, the image can become cluttered.
! A New Object-Order Ray-Casting Method – The desire is to have a high-quality, interactive volume-rendering
application. The four types of volume-rendering techniques in use today, their advantages, and their drawbacks
were addressed. The presenter took the Shear Warp approach and improved its quality while also improving its
performance by skipping empty and hidden regions. This method was then compared with the Volume Pro hardware
volume-rendering board.
The Resource, Spring 2003 ERDC MSRC 5
An HPC-Enabled Virtual Proving Ground for Seismic Unattended
Ground Sensor Networks
By Dr. Mark L. Moran, Battlefield Seismic-Acoustic Sensors Program,
ERDC Cold Regions Research and Engineering Laboratory
Comprehensive, reliable, situation information is property contrasts. Tactically significant terrain
imperative for the success of light-armor, maneuver- includes large-scale physiographic features (such as
dominated Future Combat System (FCS) operations. A forests, hills, passes, narrow valleys, or rivers), which
core thrust of the Army Science and Technology is almost axiomatic. These complex battlefield
program is directed toward developing an interlocking environments are extremely difficult sensor settings.
and overlapping network of ground, air, and national For example, a single impulsive force applied to the
asset sensor systems with the aim of delivering a earth’s surface often results in three to five distinct
timely and detailed battlefield operational picture to seismic wave fields all propagating along different ray-
commanders at all echelons. The ERDC Cold Regions paths, with different amplitudes, different decay rates
Research and Engineering Laboratory’s seismic (in both space and time-frequency dimensions),
signature simulation Challenge grant is supporting the
different polarizations, and with propagation speeds
development of a family of intelligent unattended
that vary with frequency (i.e., strongly dispersive
ground sensors (UGS) including the UGS sensors in the propagation). Moreover, all these seismic wave phases
U.S. Army Communication-Electronic Command
have complex interactions with topography and
Night Vision Laboratory’s Networked Sensors for the
geologic discontinuities (reflections, refractions,
Objective Force Advance Technology Demonstration,
diffractions, and mode conversions). However, it is by
the FCS Intelligent Munition System, and the FCS
virtue of the signal interaction with geology that results
Tactical Unattended Ground Sensor. These systems
in the over-the-hill, nonline-of-sight, sensing capabili-
rely on seismic and acoustic sensors to detect, track,
ties that are important to UGS systems.
and classify a wide range of threat targets from heavy
armored vehicles to dismounted infantry. Seismic By combining massively parallel HPC computational
signals are generated by targets via ground vibrations resources with state-of-the-art numerical methods, a
in complex ways that convey target-specific features “virtual proving ground” (VPG) has been developed for
useful for classification or identification. For example, simulating the performance of networks of seismic
the size and number of track blocks, the diameter and UGS systems in realistically complex geologies. The
separation between wheels, or the resonance frequen- resulting simulated data have been demonstrated to be
cies of the vehicles sprung mass are readily observable indistinguishable from actual field data, even by
in sensor data. Seismic signals arriving at a sensor subject matter experts. The seismic VPG capability has
also interact with geologies having large material a wide number of applications that significantly
Figure 1. Examples of detail for the M1 main battle tank mechanical model. Other targets available include
personnel, T-72, BMP-2, BTR-80, and the HMWVV. Convoys in any number or combination can also be
modeled. The detailed mechanical models generate complex distributions of target-specific ground vibrations.
6 ERDC MSRC The Resource, Spring 2003
accelerate the pace of UGS technology development, example, in complex terrains the HPC simulations have
improve system reliability, and reduce overall costs. predicted that seismic sensors will perform better
For example, using large-scale simulations, new when placed at the top of hills or in deep ravines as
methods for adapting networks of intelligent seismic opposed to flat alluvial soil deposits at the base of
UGS systems to their specific deployment environment hills. This counter intuitive conclusion is explained by
have been developed and demonstrated, providing for the HPC simulation results by noting that strong,
robust all-weather target tracking performance. As interfering signal reflections are trapped in the allu-
another example, full wave field simulations with this vium whereas the tops of hills (with stiff rock cores)
level of fidelity can be used directly for system- and deep ravines (near to water table) have geologies
specific engineering in the same manner as field data. with generally higher velocity materials, which do not
In the early stages of system development, this saves trap multiple reflected signals. HPC-supported compu-
many millions of dollars by reducing the number of tations of this nature are used to develop sensor
field studies required to develop system algorithms performance prediction maps and sensor deployment
and perform rigorous engineering trade studies to doctrine under a wide variety of geologic conditions.
select the optimal sensor suite for a given application. This is important for maximizing information quality
Lastly, the simulated data will allow analysis and and coverage area for a fixed number of networked
prediction of sensor network performance. For elements.
Figure 2. Iconic representation of the seismic virtual proving ground enabled by full exploitation of HPC
facilities. The modeling approach uses realistic heterogeneous 3-D geology, soil attenuation, and
topography as input along with signature ground vibrations. These inputs are applied to full wave
field simulations resulting in simulated data that are indistinguishable from field data.
The Resource, Spring 2003 ERDC MSRC 7