Best in class
Most opportunistic cycles for a site: Syracuse University OrangeGrid (SU-OG), which provided over 4.4 million hours.
Most efficient experiment on FermiGrid that used more than 500,000 hours since Oct. 1: Minos (98.24%) and MU2E (98%).
Most efficient big non-production user on FermiGrid since Oct. 1: Ashley M. Timmons from Minos
with 98.6% efficiency.
Experiment with the most opportunistic hours on OSG between Oct. 1 and Nov. 30: Mu2e with 11,162,407 hours.
This newsletter is brought to you by:
is the second in a series of newsletters to the community.
We welcome articles you might want to submit. Please email email@example.com.
The October newsletter is available here.
NOvA success for first results
The NOvA experiment, the largest running experiment
at Fermilab, studies the oscillation parameters that define
neutrino transformation. During the preparation for DPF, dCache was delivering
files at a rate of approximately one terabyte per hour for analysis jobs.
The utilization of FTS, SAM, and dCache allowed for complete integration
into essentially all workflows without customization by analyzers.
While data-heavy processing was focused on workers
nodes near (but not exclusively at) Fermilab, processing, such as Monte
Carlo generation, was transitioned to offsite resources. All off-site opportunistic processing
combined resulted in over 5 million CPU hours and
increased the average number of cores utilized by NOvA from 2,200 cores
on site to 3,250 cores total. More information
GENIE using OSG to improve neutrino interaction modeling
One of the goals of the Fermilab GENIE group has been to move its
validation processing to the Open Science Grid.
Photo courtesy Luanne O'Boyle.
Preparing a GENIE
physics release involves intensive computation that is not practical in a
desktop environment. The work is largely "embarassingly parallel," making it easy to spread
out over the Grid and finish in a matter of hours what might otherwise
take weeks. More information
dCache: scaling out to new heights
The dCache data storage system was first adopted by the CDF Tevatron experiment
and then became a backbone of regional CMS Tier-1 data center storage. dCache plays a very important role in helping to deliver major scientific results such as traces of the Higgs boson particle in digitized form.
The world map above shows distribution of dCache clients that have transferred at least one terabyte of data in the last three months. More information
OPOS: the importance of collaboration and cooperation
OPOS group facilitates the transfer of tools and know-how among
experiments by helping with the adoption of common tools, such as FIFE's Jobsub, SAMweb, IFDH, etc. As Tingjun Yang said, “I think the
OPOS group is doing a fantastic job, and their contribution is very
much appreciated by the DUNE collaboration.”
Jobsub is a FIFE user's doorway into running jobs on computational grids, clouds and other HPC clusters. Jobsub provides a simple-to-use, scalable and reliable job submission abstraction layer for submitting scientific workflows that run on diverse computation resources.
Since Jan. 2015, users have consumed over 10 million hours of computing cycles every month using the Jobsub infrastructure. These numbers are expected to grow even further as more experiments start taking data and progress further into their life cycle. More information
FIFE computing at European sites
In the past year, the
site outside of Fermilab providing the most computing resources to the
NOvA experiment has been the Institute of Physics of the Czech Academy
of Sciences (Fyzikální Ústav AV ČR or FZU). In the past year,
NOvA has utilized over 6 million computational hours at FZU.
When NOvA collaborators at the Joint Institute for Nuclear
Research (JINR) in Russia were interested in providing computing resources to
NOvA, FIFE and OSG staff followed the model set by FZU and set up access to a
JINR computing cluster via an OSG site. A similar setup is currently being
established at the University of Bern in Switzerland for the MicroBooNE
experiment. More information
Intensity Frontier Data Handling (IFDH) usage helpful hints
Tips and tricks for using IFDH include:
Use a cleanup call.
Make a list of files with 'ifdh ls'.
See what's going on with environment variables.
What to do when you get an error on a copy to/from DCache on-site.
output instead of copying files back.
BlueArc unmounting from GPGrid nodes - So Long, and thanks for all the files.
A long time ago, in a cluster far, far away, it was a period of rebellion against the limitations of local batch clusters. In 2009, the 3,000 cores of the GP Grid Farm were a vast improvement over the 50-core FNALU batch system, but the load has increased. The dCache storage elements deployed in 2015 can handle the current load. BlueArc cannot. We need to proceed this year with the BlueArc Unmount process, removing even GridFTP access to BlueArc data. More information
To provide feedback on any of these articles, or the FIFE notes in general, please email firstname.lastname@example.org
The complete material (for viewing offline) is available in the following formats:
iPad, Nook (epub format)
Kindle (mobi format)
PDF (Lots of white space, though, so please consider before printing)