1

DOE Office of Science

www.science.doe.gov/scidacOctober 2003

SciDAC

Scientific Discovery through Advanced Computing

Alan J. Laub

2

Introduction

• SciDAC is a $60+M/yr pilot program for a “new way of doing science”

• first Federal program to support and enable “CSE” and (terascale) computational modeling and simulation as the third pillar of science (relevant to the DOE mission) along with theory and experiment

• spans the entire Office of Science (ASCR, BES, BER, FES, HEP, NP)

• involves all DOE labs and many universities

• builds on 50 years of DOE leadership in computation and mathematical software (EISPACK, LINPACK, LAPACK, ScaLAPACK, etc.)

3

Addressing the Performance Gapthrough Software

0.1

1

10

100

1,000

2000 2004T

eraf

lop

s1996

Peak performance is skyrocketing In 1990s, peak performance increased

100x; in 2000s, it will increase 1000x

But ... Efficiency for many science applications

declined from 40-50% on the vector supercomputers of 1990s to as little as 5-10% on parallel supercomputers of today

Need research on ... Mathematical methods and algorithms that

achieve high performance on a single processor and scale to thousands of processors

More efficient programming models for massively parallel supercomputers

PerformanceGap

Peak Performance

Real Performance

4

It’s Not Only Hardware!

Updated version of chart appearing in “Grand Challenges: High performance computing and communications”, OSTP committee on physical, mathematical and Engineering Sciences, 1992.

5

SciDAC Goals

• an INTEGRATED program to:

– (1) create a new generation of scientific simulation codes that takes full advantage of the extraordinary capabilities of terascale computers

– (2) create the mathematical and computing systems software to enable scientific simulation codes to effectively and efficiently use terascale computers

– (3) create a collaboratory software environment to enable geographically distributed scientists to work effectively together as a team and to facilitate remote access, through appropriate hardware and middleware infrastructure, to facilities, data, and human resources

with the ultimate goal of advancing fundamental research in science central to the DOE mission

6

Initial Awards Focus on Software

• Scientific Applications– Climate Simulation– Computational Chemistry– Fusion – 5 projects– High Energy/Nuclear Physics (incl.

Astrophysics) – 5 projects

• Collaboratories– Four projects

• Middleware and Network Research– Six projects

• Computer Science (4 ISICs)– Scalable Systems Software– Common Component Architecture– Performance Science and

Engineering– Scientific Data Management

• Applied Mathematics (3 ISICs)– PDE Solvers/Libraries– Structured Grids / AMR– Unstructured Grids

7

Where Did the Money Go?

• In the first year, $57M was awarded (via 51 projects ranging from as little as $50K to as much as several million dollars ) in the following ways:

– about one third each to ISICs, scientific applications, and collaboratories/middleware/networks

– about one third to BES, BER, FES, HEP, NP and about two thirds to ASCR

– slightly over one half of awards to DOE labs and the balance to universities and other research institutions

8

CSE is Team-Oriented

• successful CSE usually requires teams with members and/or expertise from at least mathematics, computer science, and (several) application areas

• language and culture differences

• usual reward structures focus on the individual

• incompatible with traditional academia

• SciDAC will help break down barriers and lead by example; DOE labs are a critical asset for early success

9

Organizational Benefits

• benefits of “team-based science” are both technical and sociological

• synergistic benefits derived from interdisciplinary interactions; application scientists can now pursue more diverse and in-depth scientific explorations (e.g., Community Climate System Model)

• SciDAC teams with membership across DOE labs and with academia have enhanced cooperation across labs thereby increasing overall performance of DOE/SC

10

Successful Launch of Program

• SciDAC under way for over two years

• first PI meeting January 2002 in Washington, DC

• theme: introduction to the integrated SciDAC program; initiation of team building

• second annual PI meeting was held March 10-11, 2003 in Napa, Calif.

• theme: assessing SciDAC progress

• the SciDAC concept is working; a cultural change is emerging

• new scientific results that would not otherwise have been possible

11

Examples of Early Success

• Steve Jardin (PPPL): “… [SciDAC] is a significant factor in our productivity, comparable to that obtained by going to the next-generation computer.”

• Tony Mezzacappa (ORNL): “The SciDAC Program is making possible a whole new class of supernova simulations. I could never go back to single-investigator research.”

• Rob Ryne (LBNL): SciDAC algorithmic advancements and visualization in accelerator design enable us to “… optimize designs to reduce costs and risks and help ensure project success.”

12

Updated Overview of SciDAC

• almost 80 “two-pagers” now available on SciDAC website: www.science.doe.gov/scidac or www.osti.gov/scidac

• divided into– Basic Energy Sciences (BES)– Biological and Environmental Research (BER)– Fusion Energy Sciences (FES)– High-Energy and Nuclear Physics (HEP, NP)– Advanced Scientific Computing Research (ASCR)

o CS ISICs (Integrated Software Infrastructure Centers)o Math ISICso Collaboratorieso Networking and Middleware

13

Future SciDAC Issues

• additional high-end computing and network resources– initial SciDAC focus is on software, but new hardware is needed now– U.S. response to Japanese Earth Simulator?– potential synergistic partnerships leveraging off the success of the

SciDAC model (e.g., ITER decision and FSP)– both capability and capacity computing needs are evolving rapidly– NSTC (OSTP) HEC Revitalization Task Force (HECRTF)

• limited architectural options available in the U.S. today– science and engineering needs require architectural diversity– math and CS research will play a key role– topical or focused computing can be a cost-effective way of providing

extra computing resources

• expansion of SciDAC program– many important SC research areas (e.g., visualization, functional

genomics/proteomics) are not yet formally included in SciDAC; computational nanoscience / materials science now included as part of Nanoscale Science Research Centers