Computing Techniques Seminar TODAY: A Case Study In Public Data Release: Flight Path of Malaysia Airlines MH370 - Thursday, Jan. 29 at 1:00 p.m. in Wilson Hall Curia II
Cookie Caucus Friday, Jan. 30, 3:00 to 4:00 p.m., WH 9W, hosted by the Scientific Data Processing Solutions Group
Steve Wolbers retirement reception: After 30 years at Fermilab, Steve Wolbers is retiring to pursue a new life on the west coast, along with Heidi Schellman, who is moving to Oregon State University as head of the Physics Department. There will be a cake and coffee reception on Wednesday, Feb. 18, from noon to 2:00 p.m. on Wilson Hall 2X for everyone to stop by and wish Steve well in his retirement from the lab.
An article by Jim Amundson, Alexandru Macridin and Panagiotis Spentzouris entitled "High-Performance Computing Modeling Advances Accelerator Science for High-Energy Physics" was published in the IEEE magazine, Computing in Science & Engineering, Vol. 16, No. 6.
LDRD projects approved: Two members of SCD have had their proposed Laboratory Directed Research and Development (LDRD) projects approved by the Directorate after a recommendation from the LDRD Selection Committee and concurrence by the DOE Site Office: Ryan Rivera (Systems for Scientific Applicationsfor his proposed project, "Off-the-Shelf Data Acquisition System," and Michael Wang for his proposed project, "High Energy Physics Pattern Recognition with an Automata Processor." Committee chairman William Wester noted, "[that] many of the non-selected proposals also represent very good ideas and may be considered for re-submittal to the FY16 Call for Proposals (expected to take place in April)."
Adam Walters and Scott Neil were elected vice-president and membership director, respectively, of the Lake Michigan 7x24-Exchange Chapter. See more information
The Facility Operations 2014 Outages and Availability Report was released on January 14. This document reviews the computing availability achieved by the resources in the Feynman Computing Center, the Grid Computing Center and the Lattice Computing Center. For 2014, the availability statistics were, according to the summary, “outstanding and the best ever recorded.” Please read the report for further information (cdweb DocDB access required).
New graphics standards:
Fermilab has updated the graphic standards for the lab, providing, among other items, updated templates, logos, color palettes and email signatures, which will help to ensure that presentations and communications from Fermilab are clearly identifiable. There are also new guidelines and rules governing the use of these assets. Please visit the Graphic Standards webpage for more information.
Training requirement for technical publications - See announcement in Fermilab Today.
The CCD Authentication Services group has a test Shibboleth federated identity single-sign-on service. Contact the authentication group at firstname.lastname@example.org if you are interested in leveraging this for your
Welcome, new employees!
Krysia Jacobs - OCIO (moving from consultant to full-time)
Anna Mezzacane - SCD/Scientific Computing Services/Scientific Data Processing Solutions
(5, 10, 15 & 20+ years)
Frank Nagy - 35 years
Mike Behnke - 31 years
Rich Thompson - 31 years
Steve Wolbers - 30 years
Liz Sexton-Kennedy - 27 years
Margaret Votava - 27 years
Terry Jones - 26 years
Ron Rechenmacher - 26 years
Paul Russo - 24 years
Chih-Hao Huang - 23 years
Don Holmgren - 20 years
Jason Greskoviak - 5 years
Karin Kemp - 5 years
The Saturday Morning Physics tours of the Grid Computing Center started up again on January 24. The next ones will take place on February 7, 21 and 28.
Now playing in the FCC lobby:
Overhead-Aware-Best-Fit (OABF) Resource Allocation Algorithm for Minimizing VM Launching Overhead. Hao Wu, Shangping Ren, Steven Timm, Gabriele Garzoglio, Seo-Young Noh. MTAGS 2014 workshop co-located with SC14. Sunday, November 16.
High Performance Computing Modeling Advances Accelerator Science for High Energy Physics. Panagiotis Spentzouris. DOE Booth at SC14. Wednesday, November 19.
PDACS – A Portal for Data Analysis Services for Cosmological Simulations. Ryan Chard, Saba Sehrish, Alex Rodriguez, Ravi Madduri, Thomas D. Uram, Marc Paterno, Katrin Heitmann, Shreyas Cholia, Jim Kowalkowski, Salman Habib. Presented by Saba Sehrish. GCE14: 9th Gateway Computing Environments Workshop at SC14. Friday, November 21.
The Open Science Grid. Lothar Bauerdick. DOE Booth at SC14. Wednesday, November 19.
Desk: Catalog TASKs and notifications to end users
When a catalog task, TASKxxxxxxx, is created from an incident (INC) or requested item (RITM), all communications to the end user/customer should be done in the the INC or RITM, not in the TASK. TASKs are meant for internal work, and communication to the customer within a task is purposefully limited for this reason.
Unlike INCs or RITMs,TASKS do not include notifications to the "Caller" of the associated INC or RITM. TASK notifications are sent as follows:
Additional comments field (green):
Sends a notification to anyone in the Assigned To or Assignment Group fields and the Watch List
Work notes field (red):
No notification sent; used for internal notes
Click for large image
FermiPoint: Professional SharePoint 2013 Training Videos now available
Over 900 SharePoint training videos are now available for anyone on site to access through a license agreement with a company specializing in SharePoint training.
Most videos target SharePoint developers, administrators and power users, but some are also intended for end-users.
You can access the SharePoint training video site from the Fermilab Library website at http://ccd.fnal.gov/library/fermilabonly/sharepointindex.html.
From the CIO: Getting even more flexible
In an effort to provide greater flexibility to the workforce, Fermilab will be launching a lab-wide flexible work arrangement policy in March. This flexibility will come in three flavors: telecommuting, flextime and alternative work schedules. By-and-large, the latter two are new to both Computing and the lab.
Flextime allows for individuals to tailor their workday to better suit their lifestyles. Individuals could come in early and leave early or come in late and leave later. Alternative work schedule means that one would work say eight 9-hour days, one 8-hour day and then have the 10th day off. People could also, for instance, work two 5-hour days and three 10-hour days.
I have asked all of our managers to think about how they might implement greater flexibility into the way we work. As you know, Computing has significant operational responsibilities, so not everyone will be able to take advantage of the full suite of offerings. However I am optimistic many of you will be able to. Managers should start out cautiously as they feel their way through this new way to work. That’s ok. This increased flexibility will need to be balanced with ensuring that we can function well each and every day.
Panagiotis, Jon and I would like to pilot this project beginning March 1 and see how it works for us. We will send details about this within the next few weeks.
The goal of all this is to not only make Fermilab and Computing a more attractive place to work but to also increase productivity by giving many of our employees the ability to decide how and when they can work most effectively to contribute to our overall mission.
I think this is an important step for Fermilab to take. I am sure such a change will require some adjusting on everyone’s part -- please be patient with the system as we learn how to do this together.
Looking to the future with SAM
Robert Illingworth (left) and Stephen White (right), the two members of the SAM team who worked on transitioning SAM’s database from Oracle to PostgreSQL under Project Manager Steve Jones (not pictured).
Many who work closely with the experiments at the lab will be familiar with SAM, the data-handling system originally designed for the DZero experiment and still in use today as the system of choice for 13 Fermilab experiments, with two more currently being added. Continuity is a great advantage: many users are well-acquainted with SAM’s operations and conversant in any particular features. However, it is not surprising that the requirements of the 1990s, when the system was designed, or the 2000s, when it was updated, are not precisely those of the 2010s. Therefore, in early 2014, members of the Data Management and Applications group, including Michael Diesburg, Michael Gheith, Robert Illingworth, Arthur Kreymer, Marc Mengel and Stephen White, started a project to update SAM for the new experiments.
This project comprised two phases. First, Illingworth, with contributions from Mengel, rewrote all of SAM’s user interfaces to use web services instead of CORBA (Common Object Request Broker Architecture) as the basis of communication between the software components. Secondly, with Project Manager Steven Jones, Illingworth and White changed the database layer: instead of using Oracle, SAM is now based on PostgreSQL. “PostgreSQL being open-source gives us a lot of flexibility,” said Illingworth. “It makes it easier to deploy in other places.”
Following these changes, the team carried out thorough testing of the new infrastructure. They were helped in this endeavor by a comprehensive test suite from a previous project. This was a significant advantage, allowing the team to make their changes with confidence.
The transition—completed just before Thanksgiving, a month ahead of schedule—was a great success. And there are further benefits beyond having an updated system. “We took the opportunity to overhaul the database schema,” said Illingworth. This has streamlined the system, making it easier to respond to requests for new experiments.
Now that the updates are in place, could it be another 10 years before further changes are made? On the contrary, the recent work has meant not only the simplification of the system, but also that it is deeply understood in its new form. This makes changes easier, allowing some intrinsic flexibility for customization for individual experiments. It also makes bigger changes possible, such as, for example, if there is a need which cannot be met in the current set up. “Rather than hammer it in until it fits, it could sometimes make sense to change the underlying database,” said Illingworth.
Following this work, SAM, it seems, can look forward to a bright future as the experiment data-handling system.
Enterprise Services Operations/Web Services Administration
I joined Fermilab in October 2014 after several years of self employment. Transitioning from a home office to a lab occupied by so many skilled and passionate people has been fantastic.
As part of the Web Systems Administration group, I am responsible for keeping the central web servers that host hundreds of websites for various departments, experiments and individuals running smoothly. In addition, I work to quickly provision new websites to meet the lab's needs.
For the next year, my focus will be on migrating websites from the aging Solaris/AFS infrastructure to the new Central Web Hosting service powered by modern Linux virtual machines. This will provide significantly higher performance, stability and security for Fermilab websites for many years to come.
Though I have only been here for 3 months, I have already learned much. I look forward to continued growth from the unique challenges presented at Fermilab and learning more from the awesome people I work with.
Systems for Scientific Applications/Scientific Computing Simulation, Physics and Detector Simulation
I joined Fermilab in 1991 as a research associate on the DZero experiment, where I was involved in the data handling and the top quark cross-section determination. I subsequently moved to the Computing Division, where I had a chance to be part of and lead several groups working on mass storage, data handling (also CDF's) and computer farms, including Farms and Clustered Systems, the group that wrote products such as FBSNG, DFARM, FCP and NGOP. I also worked with an Accelerator Division group analyzing the Tevatron beam emittance.
For the last several years I have been working in the Physics and Detector Simulation group, of which I am currently the assistant group leader. The group develops and supports simulation software for the Fermilab experiments. The software includes the Geant4 simulation toolkit and physics event generators PYTHIA and GENIE. Our group members participate in larger collaborations that develop and maintain these tools.
Regarding Geant4, our team validates its hadronic physics results by comparing the simulations with the actual data. Additionally, we perform regression testing to monitor and document the evolution of the simulations. We develop and maintain the validation framework and database with the results of the aforementioned comparisons. In our work we pay special attention to the physics processes relevant to the Fermilab experiments. We also monitor Geant4 CPU performance and look for its improvement opportunities. In addition to the toolkit testing and monitoring, we contribute to its development as well as provide guidance and support regarding its usage to the Fermilab community. Another major area of our work is research and development aiming at enabling Geant4 to run on and efficiently utilize new and emerging computing hardware such us coprocessors.
My current focus within our group is Geant4 and the Mu2e experiment. I contribute to the development of the Mu2e offline simulation software, working predominantly on its Geant4-related aspects. On the Geant4 side, I participate in the monitoring of the toolkit CPU, reviewing its computational efficiency and in the code development, concentrating on the needs of Mu2e.