11
Susanne Hambrusch Division of Computing and Communication Foundations (CCF) CISE Directorate National Science Foundation June 28, 2012
Susanne Hambrusch Division of Computing and Communication Foundations (CCF) CISE Directorate National Science Foundation
June 28, 2012
CISE Organization and Core Research Programs CISE
Office of the Assistant Director
Computing and Communications
Foundations (CCF)
Algorithmic Founda1ons
Communica1on and Informa1on Founda1ons
So7ware and Hardware
Founda1ons
Computer and Network Systems (CNS)
Computer Systems Research
Networking Technology and
Systems
Information and Intelligent Systems
(IIS)
Human-‐Centered Compu1ng
Informa1on Integra1on and Informa1cs
Robust Intelligence
CISE Cross-‐CuDng Programs
CISE Organization and Core Research Programs CISE
Office of the Assistant Director
Computing and Communications
Foundations (CCF)
Algorithmic Founda1ons (AF)
Communica1on and Informa1on
Founda1ons (CIF)
So7ware and Hardware
Founda1ons (SHF)
Computer and Network Systems (CNS)
Computer Systems Research
Networking Technology and
Systems
Information and Intelligent Systems
(IIS)
Human-‐Centered Compu1ng
Informa1on Integra1on and Informa1cs
Robust Intelligence
CISE Cross-‐CuDng Programs
Other NSF representatives at the workshop …
• Tracy Kimbrel, CCF’s Deputy Division Director
• Bala Kalyanasundaram, Program Director, CCF/AF
• John Reppy, Program Director, CCF/SHF • Fen Zhao, AAAS Fellow, CCF
10
100
1,000
10,000
100,000
1,000,000
1985 1990 1995 2000 2005 2010 2015 2020Year of Introduction
Processor Performance Plateaued Around 2004
Credit: Graph reprinted with permission from The Future of Compu.ng Performance: Game Over or Next Level? (2011) by the Na1onal Academy of Sciences.
The Expectation Gap
Microprocessor Performance “Expectation Gap” over Time (1985-2020 projected)
Image Credit: USC BMES ERC
10
100
1,000
10,000
100,000
1,000,000
1985 1990 1995 2000 2005 2010 2015 2020Year of Introduction
Processor Performance Plateaued Around 2004
Credit: Graph reprinted with permission from The Future of Compu.ng Performance: Game Over or Next Level? (2011) by the Na1onal Academy of Sciences.
The Expectation Gap
Microprocessor Performance “Expectation Gap” over Time (1985-2020 projected)
Image Credit: USC BMES ERC
Parallel Alg. Research (measured in number of papers)
“In the future, all soFware must be able to exploit mul.ple processors to enter into a new virtuous cycle with successive genera.ons of parallel hardware that expands soFware capabili.es and generates new applica.ons.”
Game Over or Next Level?
The Future of Compu.ng Performance: Game Over or Next Level?, Na1onal Academy of Sciences, 2011.
NSF’s Advanced Computing Infrastructure Plan (February 2012)
hOp://www.nsf.gov/pubs/2012/nsf12051/nsf12051.pdf
ACI Strategy #1 Founda1onal research to fully exploit parallelism and concurrency through innova1ons in computa1onal models and languages, mathema1cs and sta1s1cs, algorithms, compilers, opera1ng and run-‐1me systems, middleware, so7ware tools, applica1on frameworks, virtual machines, and advanced hardware.
21st Century Computer Architecture A community white paper (May 2012)
A CCC effort led by Mark Hill, U of Wisconsin • “Because most technology and computer architecture
innovations were (intentionally) invisible to higher layers, application and other software developers could reap the benefits of this progress without engaging in it.”
• “How can we enable the 21st century infrastructure, from
sensors to clouds, adding value from performance to privacy, but without the benefit of near-perfect technology scaling?”
• Architecture as Infrastructure: Spanning Sensors to Clouds • Energy First • Technology Impacts on Architecture • Cross-Cutting Issues & Interfaces
http://cra.org/ccc/docs/init/21stcenturyarchitecturewhitepaper.pdf
Workshop on Research Directions in the Principles of Parallel Computation (June 2012) • What role should and can the algorithms/theory
community play in an effort to exploit existing and future parallelism?
• How can effective cross-layer research collaborations between architecture and algorithm/theory researchers happen?
• How about between parallel programming languages and algorithms/theory?
• What role should algorithms/theory researchers play in a redesign of the software stack?
NSF’s message to you • Get clarifications and guidance on solicitations from
Program Directors, NSF workshops, meetings, webinars • CCF needs your proposals
– one submission every 3 years is not necessarily a good strategy
• If you are funded by NSF, serve on a panel once a year • Help us identify junior faculty for panels • NSF Highlights promote your research
– If asked, please produce an effective one for a broader audience
• NSF needs researchers to serve as program directors – A great learning experience
