NCSA Headline News

NCSA's Image Spatial Data Analysis group is looking for NCSA staff with experience in the field of computer vision or related areas such as artificial intelligence, robotics, and machine learning. At times the group needs extra help on projects and would like to have a database of NCSA staff who have the relevant skills and might be interested in these opportunities as they arise. If you have worked with support vector machines, decision trees, SIFT features, or other techniques from the aforementioned areas, please email Kenton McHenry (mchenry@illinois.edu) if you would like to be considered for future projects.

NCSA is seeking two visiting research programmer in the CyberEnvironments and Technologies Directorate. The primary responsibility of these positions is to research techniques and develop software in support of various cyberinfrastructure project efforts. Part of this work has involved methods of constructing a distributed service to carry large numbers of format conversions using existing software (referred to as NCSA Polyglot). Future work will involve adding parameter support to the software servers that make up the Polyglot service, empirically estimating the resources required by any software package with a given input, and expanding a library of content to content comparison measures (NCSA Versus) which is used to measure information loss during conversions. A good number of projects involve applied computer vision, thus a background in a relevant overlapping area is preferred.

Required Education, Experience & Training

• BA/BS in computer science, engineering or related field required (advanced degrees highly preferred). Alternative degree fields will be considered if accompanied by equivalent experience (depending on nature and depth of experience as it relates to current NCSA projects and technologies)
• Demonstrated ability to carry out both research and development: Experience contributing toward research publications; Experience in software development
• Programming in Java
• Ability to clearly communicate results (both written and oral) and their importance

Preferred Experience

• Candidate-authored research publications
• Experience in computer vision or overlapping area (e.g. robotics, artificial intelligence, graphics)
• Web development experience (e.g. GWT, RESTful Web Services)
• Programming skills in other languages (e.g. C/C++)
• Familiarity with Linux and experience with HPC environments and cloud computing
• Ability to interact with people in a wide range of educational, scientific, and engineering disciplines to advance cyber-infrastructure, data curation, computer vision, high performance computing and their applications

This is a visiting academic professional position (covered by a collective bargaining agreement) at NCSA and is an annually renewable, 12/12, 100%-time appointment with regular University benefits. Salary is commensurate with experience and start date will be as soon as possible after the close date of the search. Interviews may be conducted before the closing date, although no hiring decisions will be made until after the search has closed.

To apply, please create your candidate profile at jobs.illinois.edu and upload your CV/resume by the close date, Feb. 29, 2012. Contact information for three references must be included on the application (letters maybe also be uploaded or sent to the contact below). For full consideration, candidates must complete the Hiretouch application process by the above date.

Illinois is an Affirmative Action/Equal Opportunity Employer and welcomes individuals with diverse backgrounds, experiences, and ideas who embrace and value diversity and inclusivity. For further information regarding our application procedures, you may visit www.ncsa.illinois.edu or email Jonathan Howell.

NCSA seeks a graduate research assistant to work on the Blue Waters project. This project deals with the static and dynamic optimization of communication schedules. Communication schedules are a set of communication operations and dependencies and define an order of their execution. A set of such operations and dependencies form a global communication graph. The goal of this project is to optimize the communication graph in a given model (e.g., LogGP) and to compare the quality of solutions.

For example, a broadcast communication from node 0 to nodes 1..3 can be expressed as the set {(0,1), (0,2), (0,3), (0,4)} (where a tuple (x,y) represents communication from x to y) or {(0,1}, (1,3), (0,2)} in a tree-like shape. Using a broadcast tree is more efficient in this trivial example. The project aims to develop model-based optimization techniques for the optimization of such communication operation represented in the tuple-form above. The main work is to develop and proof optimality of algorithms working on this tuple-form using well-known communication models such as LogGP.

Reaching optimality is generally very hard. We plan to follow three avenues: (1) analytical algorithms and proofs, (2) well-known optimization methods (linear or integer optimization), and (3) heuristics and learning-based methods. The results should be implemented in an MPI-like library.

For more details and references see the project webpage http://www.unixer.de/research/compi/.

Required Skills

The student working on this project should know what MPI is, be familiar with the C and C++ programming languages and should be very familiar with linear optimization, (mixed) integer programming and basic network models. The student should understand the papers Alexandrov et al. "LogGP: incorporating long messages into the LogP model—one step closer towards a realistic model for parallel computation" and Bruck et al. "Efficient Algorithms for All-to-All Communications in Multi-Port Message-Passing Systems" well.

This graduate research assistant position is a 50% appointment, starting as soon as possible, and is paid monthly. The salary is dependent on graduate status (qual/non-qual) and this position is tuition-waiver eligible. Please apply via email to Torsten Hoefler, htor@illinois.edu.

Illinois is an Affirmative Action/Equal Opportunity Employer and welcomes individuals with diverse backgrounds, experiences, and ideas who embrace and value diversity and inclusivity. For further information regarding our application procedures, you may visit www.ncsa.illinois.edu or email Jonathan Howell.

The main doors to the NCSA Building now are programmed to unlock at 7 a.m. on weekdays.

NCSA Public Affairs plans to update and improve the NCSA website this year. As we begin, we are interested in hearing from NCSA staff about what they think is most important, how the website can better serve our core audiences, and other suggestions. A feedback/brainstorming session will be held at 2 p.m. Monday, Jan. 30 in Room 1040; all staff are invited to attend. If you cannot attend, you are also welcome to send suggestions to tlbarker@ncsa.illinois.edu.

Edward Seidel, assistant director of the National Science Foundation's Directorate for Mathematical and Physical Sciences, will give a talk on The Data and Compute-Driven Transformation of Modern Science at 11 a.m. on Jan. 26 in the NCSA Auditorium. 

ABSTRACT: We all know that modern science is undergoing a profound transformation as it aims to tackle the complex problems of the 21st Century.  It is becoming highly collaborative; problems as diverse as climate change, renewable energy, or the origin of gamma-ray bursts require understanding processes that no single group or community alone has the skills to address. At the same time, after centuries of little change, compute, data, and network environments have grown by 9-12 orders of magnitude in the last few decades.  Moreover, science is not only compute-intensive but is dominated now by data-intensive methods.  This dramatic change in the culture and methodology of science will require a much more integrated and comprehensive approach to development and deployment of hardware, software, and algorithmic tools and environments supporting research, education, and increasingly collaboration across disciplines.

NCSA collaborator Franck Cappello of INRIA will give a talk at 4 p.m. Thursday (Jan. 26) in 3405 Siebel Center on "Redesigning Fault Tolerance for High Performance Computing." This talk is part of the Illinois-Intel Parallelism Center (I2PC) Distinguished Speaker Series. The seminar will be streamed live at http://media.cs.illinois.edu/live/I2PClive.asx and questions can be asked via chat at http://i2pc.cs.illinois.edu/chat.

Abstract: Fault tolerance is already a major concern for users of large-scale message passing HPC applications. Future HPC systems with their projected shorter MTBF will make this problem even more difficult to address. The current fault tolerance approach in HPC essentially relies on concepts defined 30 years ago for generic distributed systems. Several recent studies question its applicability to next generation HPC systems and applications and advocate for the exploration of novel, potentially disruptive fault tolerance techniques.
 
In this talk, we will analyze two important components of the fault tolerance design: the HPC applications and the HPC systems. We will show how we can conceive more efficient fault tolerance based on the fundamental characteristics of the HPC applications and the dynamic behaviors of the HPC systems. In particular we will present the notions of send-determinism, partial restart, hybrid fault tolerance protocols and processes clustering in HPC applications. We will also show how we can improve fault tolerance by exploring the mine of information generated dynamically by HPC systems about the state of their components.
 
Bio: Cappello holds a senior researcher position at INRIA and is a visiting research professor in the Illinois Department of Computer Science. He is the co-director (with Marc Snir) of the INRIA-Illinois Joint Laboratory for Petascale Computing. He is member of the executive committee of the International Exascale Software Project and chairs the "system software ecosystem" for the European Exascale Software Initiative. Before 2009, he led the Grid5000 project, a nationwide computer science platform for research in large-scale distributed systems used by hundreds of researchers. He is editorial board member of the international Journal on Grid Computing, Journal of Grid and Utility Computing and Journal of Cluster Computing.