GENIE (Generates Events for Neutrino Interaction Experiments) is a key part of the simulation software stack for just about every neutrino experiment at Fermilab. To simulate an accelerator-based neutrino experiment, the software is typically divided into three parts--one which handles the production of the neutrino beam, one which handles the primary interaction of a neutrino with its target and one which handles the propagation of the produced particles through the detector. GENIE is produced by an international collaboration of scientists that sits in the middle of that stack. It simulates both the dense nuclear matter of the target as well as the neutrino interaction dynamics. While GENIE has been in existence since 2004, Fermilab only recently joined the collaboration with an effort led by Gabriel Perdue that started in late 2013. One of the things that makes an institution such as Fermilab an attractive member in a collaboration like GENIE is the
laboratory's computing expertise and its ability to operate at scale.
Photo courtesy Luanne O'Boyle. The Fermilab group consists of (from left to right), Robert Hatcher, Julia Yarba, Gabriel Perdue and Tomasz Golan, members of the Fermilab Physics and Detector Simulation group.
From the beginning, one of the goals of the Fermilab GENIE group has been to move its validation processing to the Open Science Grid. Preparing a GENIE physics release involves intensive computation that is not practical in a desktop environment. For example, cross sections as a function of energy need to be computed on dozens of materials for almost a hundred different interaction models, and some models take more than a day to compute. But 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. FIFE tools, Jobsub in particular but also data storage and data handling tools like IFDHCP, make it straightforward to organize the complex production of these calculations.In addition to total cross-section computations, GENIE recently began running large simulations to study the propagation of daughter particles from neutrino interactions through the nucleus. Very large numbers of events are required to study these simulations in detail, and Grid-scale computing is an important enabling technology. Our goal is to move more and more of the validation onto the Grid and automate the production of validation samples in order to free GENIE developers to focus on improving the physics of the simulation instead of organizing production. FIFE team members Neha Sharma, Ken Herner, and Tanya Levshina have been particularly helpful in our pursuit of this goal--so thanks!
- Gabriel Perdue