Service

Service Statement

Since coming to Boulder in 2002, I have actively participated in a wide spectrum of service activities that have directly benefited CU, NCAR, and the broader computational science community. Below I elaborate on my efforts to support computational science programs at CU and NCAR, graduate education, and highlight what I believe are my highest impact community service activities. A complete listing appears in my CV.


1. University Computational Science Program

I have served in a variety of capacities in support of the creation of an active computational science initiative at CU. In 2008, I served on the Interdisciplinary Computational Science and Engineering (ICSE) Steering Committee created by the Vice Chancellor for Research (VCR) Stein Sture and chaired by Prof. Mark Ablowitz. The committee met during the spring and early fall of 2008 and submitted its report and recommendations to the VCR on October 2, 2008. The report is congruent with the CU Flagship 2030 Strategic Plan and recommends beginning a broad effort in ICSE, including a graduate degree-granting unit with faculty (ten new faculty lines over six years), research assistants, and funding for graduate student teaching, research assistantships, and three-year postdoctoral/instructorships.


Concurrent with this effort, VCR Sture encouraged me to think about how the campus could better position itself to be able to satisfy the anticipated high-performance computing needs of such an initiative. I suggested that we build off of our first (and highly successful) collaborative NSF MRI grant and submit a follow-on proposal for a supercomputing system to help the campus address the technical and practical obstacles to practical petascale computing for a wide range of researchers in areas aligned with the campus’ research efforts in earth system science, biotechnology / life sciences, and energy / renewable energy. The resulting NSF MRI proposal (a consortium of three institutions involving 62 senior personnel) was funded in June 2008 and thus began a long series of discussions on how to deploy and support the system. The technical details of Janus supercomputer are discussed in my research statement.


In the summer of 2009, I was asked to serve as the Faculty Director of Research Computing by Interim VCR Russ Moore. The objective of this appointment was to provide the authority to act on behalf of the campus to deploy the system and advise the campus on how to structure research computing and better coordinate future cyberinfrastructure investments. To broaden the scope of input on these questions, the Research Computing (RC)/Cyberinfrastructure Committee was formed in the spring of 2010, chaired by ITS Director David Bodnar and Interim VCR Russ Moore, and on which I served. The result of this was to establish Research Computing as a formal entity reporting to the CIO and VCR’s offices and the creation of Boulder Campus Cyberinfrastructure Board. I drafted the charters for both entities and serve as a member of the latter. I believe that my contributions to the campus are best summarized by the input provided to my Department by (then Interim) Provost Russ Moore:

Let me state that Henry certainly put in yeoman's work over the last several years (last year in particular) to on campus service that culminated in the acquisition of a research computing resource for CU Boulder that ranked #31 on the most recent top 500 list. From my perspective I would certainly argue that Henry's service contributions were indeed substantial (incredible actually), and that yielded a resource for the whole CU Boulder campus that promises to put us in the top echelon (from a hardware perspective) of universities (or any organization for that matter). While this resource is indeed a campus resource, those in the natural sciences and engineering will be the largest and most immediate beneficiaries.


The work is not done. Now that the resource has been designed and acquired (again, largely through the brute force efforts of Henry), we have more work to do to `stand the system up' and staff it. We are doing so, and Henry will again give of his time to provide guidance and insights into this process working into the next AY. Certainly from my perspective, I think his service contributions to the campus have been nothing short of heroic.

Overall, this experience provided an opportunity for me to serve the University by helping formulate and implement its long-term plan to provide computational resources to its diverse research community. Originally articulated as early as 2003 as building a "watering hole" for high-performance computing research and education (with Profs. Oliver McBryan and Elizabeth Jessup), this work has expanded in scope during my tenure at the University of Colorado from serving my research group, the Computer Science department, and the College of Engineering, now to the entire campus with support from the Vice Chancellor for Research, Provost, and others in the Administration.


2. NCAR Computer Science Section

The most significant professional service I have done is as a member of the Computer Science Section in the Technology Development Division at NCAR. Over the past eight years, I have taken on an increasing set of responsibilities. During the first year and a half, my primary responsibility was to help organize and set the Section’s research direction. In the spring of 2004, I was asked to serve as Computer Science Group Head, and in August of 2005 I became interim Numerical Methods Group Head. In 2006, the Section was reorganized and I was appointed manager. Since that time I have recruited all but one of the current members of the Section and changed the composition from one dominated by software engineers to one with an equal mix of software engineers, ladder-track scientists, and student assistants. I built a research-computing environment using the Section’s equipment budget and outside funding and formed the Research Systems Evaluation Team (ReSET), which is responsible for technology tracking, managing our experimental cyberinfrastructure, and TeraGrid operations. I directly manage all of the staff, perform annual reviews, set raises, and am responsible for initiating promotions. Given the two positions at NCAR and CU, I effectively manage a small department with twelve to fifteen staff and an annual budget of around three million dollars.


3. Computer Science Department Service

I have served on my Department's Graduate Committee since coming to CU, save for the 2006-2007 academic year, which was my first year managing the Computer Science Section at NCAR. I joined because I needed graduate students and remain because I firmly believe that for the department to excel it must have a vibrant and strong research program, for which excellent graduate students are a necessary component. Under the leadership of Professors Cai, Gabow, and Mishra, the chairs during my tenure, and working with the other members of the committee, I have been involved in every aspect of the Graduate Committee’s business: M.E., M.S., and Ph.D. admissions; writing letters of support for the Dean's and Chancellor's fellowships; annual graduate student progress evaluation; TA assignment, evaluation, and awards; grading M.S. comprehensive exams; hearing student petitions; reviewing, approving, and scheduling area exams; reviewing and approving new graduate courses; travel awards; and redesigning the preliminary examination system. Once challenge that I’ve taken on personally is graduate student recruiting. The challenge the Department faces is how to recruit the best students possible given a declining number of applications and a less than favorable funding environment. To help the Department increase its yield of highly competitive students, I have organized and run the graduate student recruitment weekend five times, with Profs. Doug Sicker, Rich Han, and Evan Chang. Many of our current students decided to come to Boulder because these efforts made the difference in their decision making process.


4. Community Service Highlights

I have been a reviewer for numerous conferences and journals and have served on program committees for wide range of conferences. My most important contribution in this area is my long association with the LCI International Conference on High-Performance Clustered Computing. This conference is the premier international forum to share management, administration, and scientific computing expertise and experiences on clusters at all scales. This is a four- day event including keynotes, tutorials, and talks featuring a broad range of material and papers from high-performance computing and large-scale cluster computing professionals in industry, academia, and government. I served in the program committee for the 2005, 2006, and 2007 conferences where my primary duty was to review papers. For the 2008 conference I served on the steering committee and as technical program chair, where I significantly expanded the membership of the program committee, created a new set of advertising outlets (e.g., I negotiated a co-marketing agreement with HPCwire), established a student travel award program, and increased the amount of money awarded for best student papers and posters. For the 2009 conference I was elected conference chair and hosted the event in Boulder at NCAR. The conference theme examined how future large-scale system design will address the balance between inter- and intra-node parallelism and how applications will effectively harness these systems. I was responsible for all conference planning and logistics and raised the majority of the sponsorship money, which was problematic due to the extremely poor economic environment. The conference attracted 135 attendees from around the world, with the majority coming from the United States; attendee feedback was uniformly positive.


My expertise is widely sought after. I'm a member of Microsoft's Technical Computing Executive Advisory Committee (http://www.modelingtheworld.com) providing advice on technical computing tools and techniques, parallel program development, and cloud computing. I serve on the TeraGrid Resource Allocation Committee (TRAC - www.teragrid.org), which is responsible for allocating all of NSF’s cyberinfrastructure resources (e.g., computing cycles, storage, visualization, consulting, etc.) to computational researchers in support of NSF’s research mission. Finally, I am a member of the GECO (Golden Energy Computing Organization - geco.mines.edu) advisory committee providing advice on research computing, high-performance computing equipment, and allocations.