Computational Research Division CRD Strategic Outlook

1
Computational Research Division (CRD)Strategic
Outlook

• Horst D. Simon
• March 26, 2003

2
Computational Science Mission
The Computational Research Division is engaged in
computational science collaborations, creating
tools and techniques that will enable
computational modeling and simulation, and lead
to new understanding in areas such as
3
Computer Science and Applied Mathematics Mission
The Computational Research Division is engaged in
basic and applied research addressing the
following questions of fundamental importance to
enabling progress in our ability to use computing
and networking technology
4
Computational Research Division

• Horst Simon, Division Director
• The Computational Research Division (CRD)
  creates computational tools and techniques that
  enable scientific breakthroughs, by conducting
  applied research and development in computer
  science, computational science, and applied
  mathematics. CRD consists of two departments
• High Performance Computing Research
• Distributed Systems
• About 5M of SciDAC funded projects

5
High Performance Computing Research Department
(HPCRD)

• Juan Meza, Department Head
• Groups
• Applied Numerical Algorithms
• Center for Computational Sciences and Engineering
• Future Technologies
• Imaging and Informatics
• Scientific Computing
• Scientific Data Management
• Visualization
• Total Staff 108

The High Performance Computing Research
Department conducts research and development in
mathematical modeling, algorithmic design,
software implementation, and system
architectures, and evaluates new and promising
technologies.
6
Distributed Systems Department

• William Johnston, Department Head
• Deb Agarwal, Department Deputy
• Groups
• Collaboration Technologies
• Data Intensive Distributed Computing
• Network Technologies
• Secure Grid Technologies
• Total Staff 25

The Distributed Systems Department researches
and develops software components that allow
scientists to address complex and large-scale
computing and data analysis problems in a
distributed environment such as the DOE Science
Grid.
7
Programmatic Funding in Computing Sciences (FY02
in M)
8
Berkeley Lab is Largest Recipient of SciDAC
Funding
- new FY2002 funding to CRD 5.3M - 39 of all
lab funding

1
2
ANL
11
3
1
BNL
LBNL
LLNL
36
LANL
ORNL
39
PNNL
1
PPPL
Sandia
6
9
LBNL SciDAC Portfolio Projects on Which LBNL is
the Lead Laboratory
10
LBNL SciDAC Portfolio LBNL as a Collaborator on
Projects Lead Elsewhere
11
Existing CRD Collaborations

• By its very nature computational research
  collaborates with virtual all scientific
  divisions at Berkeley Lab
• Tight linkage with NERSC Center (10 FTE matrixed)
• Close collaborations between UCB and UCD and CRD
  in computer science and computational science
• 11 joint faculty positions
• Major projects in visualization and parallel
  programming language in CRD
• Collaborations with almost all SC labs through
  SciDAC
• Collaboration with LLNL (ASCI platforms)
• Collaborations with NASA in grids, astrophysics,
  and AMR methods

12
Future Projects

• Ultrascale Computing
• Will support architecture, tools, algorithms
• More details in the NERSC presentation
• Applied Math/Nano Science
• Two joint proposals with MSD and UCB in progress
• Cosmic Simulator
• With PD, NSD develop a complete model of the
  history of the universe
• SNAP Computing
• Computing is strategic part of this new facility
• Established SNAP Computing activity under Bill
  Johnston

13
Future Projects (cont.)

• Computational Biology and Genomics
• Strengthen and grow Genome to Life
  collaborations with PBD
• Built on visualization collaboration with JGI

14
Positive Strategic Outlook for CRD

• Computational modeling and simulation is
  increasingly being recognized as the third leg of
  science
• High priority in Office of Science
• Berkeley Lab and UC provide a stimulating
  environment with many exciting new collaboration
  activities
• Supportive lab environment for computing
• SciDAC is an intellectual investment with great
  opportunities to leverage from
• Presence of NERSC continues to be a great benefit
  for CRD
• Use of facility
• New, stimulating computer science questions
• Strong support from all OASCR program managers

15
Challenges and Issues

• Base budget in math and CS is flat
• Further growth of the OASCR program depends on
  Ultrascale initiative (USSCC)
• SciDAC has a three to five year window
• Midterm (18 months) reviews of CRD projects in
  2003 went very well so far
• But no long term decision on future of SciDAC yet
• CRD is scattered across Bldg 50 complex need
  better space to build collaborations (50X)