# InTheLoop | 09.08.2014

## Bell Named to National Academy of Sciences Board on Mathematics

**John Bell**, a senior staff mathematician in CRD’s Center for Computational Sciences and Engineering (CCSE), has been named to the Board on Mathematical Sciences and their Applications (BMSA) of the National Academies of Sciences. Established in November 1984, the board leads activities in the mathematical sciences at the National Research Council (NRC).

“This is yet another well-deserved honor for John and a fitting recognition of his leadership in innovative applied mathematics,” said Computational Research Division Director David Brown.

Bell, who recently stepped down as leader of CCSE after more than 20 years, is well-known for his work in the areas of finite volume methods, numerical methods for low-Mach-number flows, adaptive mesh refinement and parallel computing. He has also worked on the application of these numerical methods to problems from a broad range of fields, including combustion, shock physics, seismology, flow in porous media and astrophysics. More recently, he has been developing new methods for combining experimental data and simulation to provide better scientific predictions.

He was the recipient of the SIAM/Association for Computing Machinery (ACM) Prize in Computational Science and Engineering in 2003, and he received the Sidney Fernbach Award in 2005 “for outstanding contributions to the development of numerical algorithms, mathematical and computational tools and on the application of those methods to conduct leading-edge scientific investigations in combustion, fluid dynamics and condensed matter.”

The BMSA consists of 19 members whose backgrounds represent the wide range of the mathematical sciences: core mathematics, applied mathematics, statistics, operations research, scientific computing, and financial and risk analysis. This composition of the board reflects a deep conviction of the unity of the mathematical sciences in all their manifestations and of the importance of adding to both the knowledge in and the applicability of the mathematical sciences.

## Calculations Confirm Properties of New Catalyst that Converts CO2 to Fuel

Scientists from the University of Illinois at Chicago have synthesized a catalyst that improves their system for converting waste carbon dioxide (CO₂) into syngas, a precursor of gasoline and other energy-rich products, bringing the process closer to commercial viability.

Supercomputing resources at the U.S. Department of Energy’s National Energy Research Scientific Computing Center (NERSC) helped the research team confirm their experimental findings. »Read more.

## Manycore Code Optimization Tool Subject of Upcoming NERSC Webinar

NERSC is hosting a webinar about the Cray Reveal tool for porting to manycore systems. Heidi Poxon, technical lead and manager of Cray Performance Tools, will present the hour-long web training at 10 a.m. PDT, Thursday, September 18. Cray Reveal assists users in optimizing code by providing variable scoping feedback and suggested OpenMP compiler directives. NERSC users considering adding OpenMP to their applications will find this training especially interesting. »Read more.

## This Week's CS Seminars

### UCB Simons Institute Open Lecture: Geometry, Invariants and the Elusive Search for Complexity Lower Bounds

#### Monday, Sept. 8, 4 p.m. - 5 p.m. (refreshments served at 3:30 p.m.)

Banatao Auditorium, Sutardja Dai Hall - UC Berkeley Campus

#### Peter Bürgisser, TU Berlin, Germany

Since the 1970s, it has been known that algebraic geometry provides concepts and methods for proving that the evaluation of certain polynomials is hard. However, the range of these methods so far has been limited, and its successful application to algebraic versions of NP-complete problems (e.g., computing the permanent) has remained elusive. The approach of attacking these problems using the representation theory of groups (symmetries) reveals fascinating and unexpected connections with other areas of mathematics. A remarkable recent insight is an intimate connection between effective questions of classical invariant theory (as studied in the 19th century) and the presumed hardness of the permanent.

This talk will survey these connections at a high level. No detailed knowledge of algebra or complexity theory will be assumed. »Learn more.

### Bayesian Numerical Homogenization

**Wednesday, Sept. 10, 2014, 3:30 p.m. - 4:30 p.m. , 939 Evans Hall **- UC Berkeley Campus

**Houman Owhadi, CalTech**

Numerical homogenization, i.e., the finite-dimensional approximation of solution spaces of PDEs with arbitrary rough coefficients, requires the identification of accurate basis elements. These basis elements are oftentimes found after a laborious process of scientific investigation and plain guesswork. Can this identification problem be facilitated? Is there a general recipe/decision framework for guiding the design of basis elements? We suggest that the answer to the above questions could be positive based on the reformulation of numerical homogenization as a Bayesian Inference problem in which a given PDE with rough coefficients (or multi-scale operator) is excited with noise (random right hand side/source term) and one tries to estimate the value of the solution at a given point based on a finite number of observations. We apply this reformulation to the identification of bases for the numerical homogenization of arbitrary integro-differential equations and show that these bases have optimal recovery properties. In particular we show how Rough Polyharmonic Splines can be re-discovered as the optimal solution of a Gaussian filtering problem.

### Exascale Node Technologies Overview: DOE's Fast Forward Projects (AMD, IBM, Intel, NVIDIA)

#### Friday, Sept. 12, 11 a.m. - 12:30 p.m., Bldg. 50B, Room 4205

John Shalf and Nick Wright, Lawrence Berkeley National Lab

We will present the latest progress on DOE's Fast Forward project, which is making investments in advanced node technologies for exascale computing.

The objective of the FastForward program is to initiate partnerships with multiple companies to accelerate the R&D of critical component technologies needed for extreme-scale computing. It is recognized that the broader computing market will drive innovation in a direction that may not meet DOE’s mission needs. Many DOE applications place extreme requirements on computations, data movement, and reliability. FastForward has funded innovative new accelerated R&D of technologies targeted for productization in the 5–10 year timeframe. The period of performance for the FastForward program is two years, with the intent to establish an ongoing program for successful projects to continue innovation in these and additional technology areas. In 2011, the FastForward RFP solicited innovative R&D proposals in the areas of processor, memory, and storage, and I/O that will maximize energy and concurrency efficiency while increasing the performance, productivity, and reliability of key DOE extreme-scale applications from both the NNSA and the Office of Science. The goal is to begin addressing long-lead time items that will impact extreme-scale HPC systems later this decade. Technology roadmaps, as they exist today, threaten to have a hugely disruptive and costly impact on development of HPC applications and ultimately a negative impact on the productivity of DOE scientists.

While DOE’s extreme-scale computer requirements are a driving factor, these projects also demonstrate the potential for technology adoption by broader segments of the market outside of DOE supercomputer installations. This public-private partnership between industry and the DOE, initiated with FastForward, will aid the development of technology that reduces economic and manufacturing barriers to constructing exaflop-sustained systems, but also further DOE’s goal that the selected technologies have the potential to impact low-power embedded, cloud/datacenter, and midrange HPC applications. This ensures that DOE’s investment furthers a sustainable software/hardware ecosystem supported by applications across the HPC market and the broader IT industry. FastForward investments will result in an increase in DOE’s ability to leverage commercial developments for future systems.

REMOTE ACCESS INSTRUCTIONS for AUDIO & WEB CONFERENCE

To begin AUDIO conference:

- Dial Toll-Free Number: 866-740-1260 (U.S. & Canada); International participants dial:
- Toll Number: (303) 248-0285, or
- Toll-Free: http://www.readytalk.com/intl

- Enter 7-digit access code: 4864508, followed by “#”
- Place your phone on "Mute"

To begin WEB conference:

- Go to:http://www.readytalk.com
- Enter your access code: 4864508 under “PARTICIPANT-Join A Conference”