CISL Director's Message

Welcome to the FY2016 CISL Annual Report from Anke Kamrath, CISL’s Interim Director.

It is my sad duty to report that NCAR lost a beloved and respected colleague when Al Kellie passed away unexpectedly in his sleep on 7 September 2016.

Al Kellie
Al Kellie, SCD and CISL Director 1998-2016.

Al will be remembered for his considerable contributions to NCAR – especially his forward-looking stewardship of its high-performance computing resources – and for the way he truly cared for his staff both professionally and personally. Al joined NCAR in 1998 as Director of the Scientific Computing Division (now CISL). Under his leadership, NCAR transitioned from a parallel vector supercomputing environment to massively parallel clusters of microprocessors, delivering a staggering 30,000-fold increase in sustained computing capacity to the atmospheric and related sciences during his tenure.

In 2003 Al spearheaded NCAR’s effort to design, fund, and construct a new high performance computing (HPC) center to support future generations of power-hungry, high-performance supercomputers. By 2009 this effort had matured into a successful proposal to the NSF for constructing the NCAR-Wyoming Supercomputing Center (NWSC) in Cheyenne, Wyoming. The NWSC became a case study in the international HPC community for its exemplary design and construction. Al was extremely proud of the award-winning NWSC when it opened in 2012, and his pride grew as CISL staff developed it and the research community used its resources to more fully realize its potential.

Anke Kamrath
Anke Kamrath, CISL Interim Director

It is my honor to serve as CISL’s Interim Director and to continue our forward-looking work to meet new challenges. As a brief introduction, my name is Anke Kamrath and I joined CISL in July 2009. Since that time I have been a member of CISL’s executive leadership team, working to support the Lab’s vision for and excellence in serving the research community. I will continue serving in an interim capacity until a permanent CISL Laboratory Director is appointed.

I am now supporting NCAR and UCAR in three major roles. I will continue working as CISL’s Director of the Operations and Services Division, and I will both oversee the Lab as a whole and serve as the Security Chief Information Officer (SCIO) for UCAR. Part of my work as SCIO will be to manage UCAR’s progress toward full compliance with the Federal Information Security Management Act (FISMA).

My first priority is to move NCAR’s newly installed supercomputer named Cheyenne into full production by January 2017. I am committed to continuing Al Kellie’s series of successes by ensuring that the NWSC’s second supercomputer is operational and advancing our science right on schedule. This annual report highlights some of the successes CISL has achieved during Al’s final year as its Director. It is organized into three sections that begin with our Service achievements in cyberinfrastructure, user support, big data, and information technology. The Science section follows with reports on our data-focused research, numerical methods for modeling, and applied computational science research. Finally, CISL’s Education efforts are described in the areas of integrating education into our research, training researchers in scientific computing, and reaching out to attract new and diverse talent to our scientific and technical enterprises.

Service

CISL’s service highlight occurred in Cheyenne, Wyoming, where the NWSC facility was fully prepared and the Cheyenne system was installed and powered up before the end of the fiscal year, all while the Yellowstone system operated at full productivity without interruption. Our operational supercomputing environment includes the Yellowstone computer, the central disk storage system known as NCAR’s GLobally Accessible Data Environment, two data analysis and visualization systems, the High Performance Storage System data archive, high-speed wide-area networking, and various high-performance data transfer and sharing services, all coupled by the many system and application software suites that allow our science to advance at an ever-faster pace. To prepare for the next generation of computing, CISL continued exploring new types of systems by operating and enhancing its HPC Futures Lab to assess new technologies that may be available in upcoming production computers. CISL’s end-to-end workflow support for Earth System scientists includes comprehensive, targeted user support, an expansive and growing portfolio of Big Data services, enterprise IT support for all of NCAR and UCAR, and leadership in cyberinfrastructure initiatives from local to international levels. CISL staff produced significant progress in these areas, and overviews of this work appear in the first section of this annual report.

Science

CISL’s science portfolio made numerous advances this year by combining petascale supercomputing resources with the latest computational science research in algorithms, mathematical techniques, and statistical methods applied to Earth System science. New scalable algorithms for data assimilation are being developed and used to support large atmosphere and ocean prediction models in collaboration with scientists throughout NCAR. Strides in interpreting and using heterogeneous data continued throughout the year. Computing for climate models was accelerated through new algorithms and by exploiting emerging technologies such as coprocessors. CISL scientists also developed computationally efficient parallelizable data analysis techniques and data processing tools. And long-term research efforts are now succeeding in building uncertainty measures into climate models in ways that are useful for decision making and policy.

Education

CISL’s education accomplishments include strong participation by new interns, students, and visiting scientists in workshops, seminars, and training programs that supplement education efforts throughout UCAR and at other institutions. Educational events continued to bring mathematicians and computer scientists to NCAR, and these led to collaborations, influenced graduate student research, and enriched our staff. Training in using HPC systems and CISL-developed data analysis and visualization tools developed more capable and effective researchers. And community outreach increased NCAR’s connection with and maximized our impact on the university community, students, interns, and future employees. CISL’s outreach efforts actively attracted qualified candidates – particularly those from diverse backgrounds – to our research enterprise and allowed us to enrich our workforce while increasing its diversity.

As you read about CISL’s accomplishments, I hope you can appreciate the importance of the progress we are making and the pride we share in our work and the dedication of our staff.