Sector news

--- Calendar migration this month. As reported in Fermilab Today, the next phase of the FermiMail migration project--calendar migration--beings next week. On Tuesday, March 20, all CS conference rooms will move from Meeting Maker to FermiMail.

Meeting Maker users received an email from the Service Desk with detailed information about the calendar migration, including an offer to opt in to the service to have their individual calendars migrated to FermiMail by the email team. The opt-in offer, which is available during next week's migration from March 21-23, is required by end of business on Monday, March 19.

Service Desk interface change: Those requesting help through the Service Desk website will soon have a slightly different interface. The new interface will have everything the current interface has, plus additional views of information, including the expanded service catalog that service providers see. See the Service Desk Tour Guide for more information.

--- Vicky White gave an opening speech at the International Symposium on Grids & Clouds in Taiwan on Feb. 28. [View all presentations by CS members]

--- CS helps EDIT: Several members of the Scientific Computing Division were engaged in the Excellence in Detectors and Instrumentation Technologies (EDIT) 2012 hands-on school held at Fermilab Feb. 13 to 24.

Gustavo Cancelo co-taught a class and lab on charge-coupled devices; Simon Kwan acted as teaching assistant in the silicon tracking lab; Lorenzo Uplegger and student Jennifer Ngadiuba ran lab classes on using the CAPTAN system and pixel readout; Andrew Norman led a tour of the NOvA Surface Building.

Chris Stoughton helped teach the Fundamentals of Radiation Detection lab. "We showed students how to use detectors with organic scintillators (plastic and liquid), inorganic scintillators (NaI), and Germanium detectors," says Stoughton. "Students used LabVIEW software on National Instruments computers to see how pulses change with particle identification and measure energy resolution, the speed of light and the muon decay lifetime."

Milestones
Person on mountain summit

Congratulations!

Maura Porta and Gabriele Garzoglio (Grid & Cloud Computing) are the proud parents of a baby girl, Isabella Maryse. Baby came into the world on Wednesday, Feb. 1 at 9:57 p.m. weighing 6 lbs., 10 oz. and 18 3/4 in. long.

Welcome, New Employee

Hyun Woo Kim (Grid & Cloud Computing/ Distributed Offline Computing Services)

Happy Anniversary!
(5, 10, 15 & 20+ years)

Ruth Pordes - 39 years
Irwin Gaines - 37 years
Ken Treptow - 34 years
Leo Michelotti - 32 years
Bill Boroski - 31 years
Mark Bowden - 29 years
Margherita Vittone Wiersma - 27 years
G.P. Yeh - 27 years
Peter Cooper - 24 years
Chuck Andrews - 23 years
Karen Shepelak - 23 years
Mike Diesburg - 22 years
Lynn Garren - 22 years
Bruce Karrels - 22 years
Bob Tschirhart - 22 years
Matt Arena - 20 years
Scott Nolan - 15 years

Noise reading update
man plugging his ears

Here’s an update to my report in August about noise levels in sector facilities. Since then, we’ve taken additional samples, and, due to the addition of equipment in some areas, the noise levels have changed.

Fermilab has adopted the recommendation by the American Conference of Government Industrial Hygienists: an eight-hour time-weighted average of 85 decibels (dBA). This means that a worker may be exposed to noise levels up to 85 dBA for an entire eight-hour shift without wearing hearing protection.

Current noise levels in FCC, GCC and LCC are available from the sector’s ES&H page or directly from DocDB. GCC Computer Room C and the FCC3 computer room have areas with noise levels above 85 dBA; there are some additional areas, indicated on the maps within each report, that are also at or above this level.

Certainly, we recommend that you wear earplugs in these rooms when you work in them for an extended period of time.  Earplugs are available near the entrances of these rooms.

~Amy Pavnica

Artful processing
Computing Center image courtesy Reidar Hahn

As computing hardware providers continue both to supply us with higher and higher core counts and to offer specialized high performance accelerators, we must develop new programming techniques so we can enhance our software to take better advantage of the parallel processing resources they offer. To this end, members of my group, Computing Enabling Technologies (CET) in the Accelerator and Detector Simulations and Support department, are working to transform the art event-processing framework into a high performance computing toolkit to better utilize readily-available multi-core and many-core machines.

Art is derived from the CMS software framework and has been restructured to work for other experiments as well. NOvA, the LArSoft Program, mu2e and g-2 are all currently using it. Upcoming experiments can directly benefit from a high performance toolkit within their real-time data acquisition (DAQ) and event filtering systems. In fact, the use cases chosen to focus and drive this project come directly from future needs of mu2e, g-2, DarkSide-50 and NOvA.

As a result of the real-time and high throughput needs of these experiments’ DAQ systems, the complete toolkit includes a distributed event building layer, which prepares the data on which art operates, along with the high performance art layer, which analyzes the data using physics algorithms. The event building layer uses the Message Passing Interface (MPI) to communicate over an Infiniband network, and OpenMP handles event-level parallel processing within art.

We have successfully stood up a development cluster with a high performance network, which allows us to feed data into the demonstrator applications at realistic rates—those necessary to keep modern 32-core servers busy. The excellent multithreading capabilities of the new C++11x library standard already available in GCC enable us to easily integrate the event building layer with art.

Our current demonstrator is configured to emulate the generation of five DarkSide-50 raw data fragment streams using three nodes, all feeding a single event-building node. The builder forms 6MB events and feeds them into five art processing channels available on that node. Early tests show that we are able to feed events to art at an overall rate of about 250Hz (rate > 1.5GB/sec). We are moving towards adding realistic processing algorithms to the art channels to better estimate the processing capabilities of the machines we have in house.

~ Jim Kowalkowski

Core Computing Spotlight

Keith Coiley
Core Computing/Facility Operations/Facilities Support Services

Keith Coiley

As a member of the Facilities Support Services team, I contribute to the efficiency of Facility Operations. My role encompasses different activities and projects as well as maintenance and logistic services within the sector’s data centers, buildings, offices and support spaces. Ensuring safety in our environment is crucial to all of these activities, so I also serve on FESHCom, the lab’s safety committee responsible for reviewing safety and security policies and programs and reporting findings and recommendations to the Fermilab director.

My work focuses on two major areas: office relocation support and task management. Once approved, I coordinate a variety of services needed for office moves, including installation of whiteboards, bookshelves, computer lockdown kits, keyboard trays, etc., relocation of phone extensions, ordering nameplates, specialized cleanings and many other office-related maintenance.

As a task manager, I also coordinate and manage laboratory staff or sub-contractors that perform work in sector facilities. Such work includes delivery and installation of large computer systems, relocation of computing equipment, biannual data center cleanings, maintenance of various types of building infrastructure, reconfiguration of office and support spaces and safety enhancements.

All these services are in place to provide a strong foundation so that Computing Sector can carry out its mission successfully.

Tips of the month

--- Service Desk


Service providers may have noticed the new “Work Notes" field in Service-Now with its nifty red background. This field enables service providers to document technical details without prompting a notification to customers or watch-list participants. The purpose of this field is to provide technical information or details the customer does not need to be concerned with.

Here are some (good and bad) examples of how to use this work notes:

Good use of Work Notes

Bad use of Work Notes

To keep a log of work completed

To keep a diary of how frustrating a particular service, product or customer is

To note details the customer isn't necessarily interested in

To note passwords or other sensitive information

To communicate with another service provider about work related to the ticket or a related ticket that the customer does not need to know

To see if your supervisors or coworkers are spying on you, i.e., “I’m glad it’s my birthday… sure would be nice to be treated to lunch today!”


--- SharePoint

The folder with the up-arrow icon SharePoint Folder icon to go up a directory moves you up one site. Did you know that clicking once shows you where you're at within the site?

You can also click on a site or subsite name to navigate throughout the site: