Editor’s note: This is an adapted excerpt from an article published by Lawrence Livermore National Laboratory. Fermilab scientists contributed to the award-winning Spack software package management tool, whose initial development was led by Livermore.
In November, the trade journal R&D World Magazine announced the winners of the annual R&D 100 Awards, sometimes called the “Oscars of invention,” in an online post and emails to winners.
Among the winners was Spack, a software package management tool for high-performance computing applications, a project led by Lawrence Livermore National Laboratory, a DOE national laboratory.
Spack is an easy-to-use, versatile and scalable software package management tool for high-performance computing applications. It simplifies and accelerates building and customizing software by automating the build workflow, thus reducing deployment time for large software stacks from weeks to hours.
In 2013, Todd Gamblin, a computer scientist at Lawrence Livermore National Laboratory, created the first prototype of Spack to automate the many tedious software builds he and colleagues were forced to perform manually.
More than 2,300 monthly active users use Spack software worldwide. Image courtesy of Spack
Today, Spack is widely available as open-source software. Spack’s original 100 or so packages have grown into a library of more than 3,500, and a large and active community of more than 450 contributors regularly adds features and improvements.
Spack is now used for software deployment on six of the world’s top 10 supercomputers, and it has been adopted by a number of high-performance computing centers and software development communities.
Even outside of high-performance computing centers, Spack has had a major impact. It has been picked as an end-to-end development tool within the high-energy physics community, based at Fermilab in the U.S. and CERN in Switzerland.
Fermilab scientists contributed to Spack’s core functionality as well as software definitions specific to high-energy physics.
While Livermore led the initial development of Spack, the software’s co-developers include Fermilab and Argonne National Laboratory, the Center for Climate Systems Research, Columbia University, Iowa State University, Kitware Inc., NASA Goddard Institute for Space Studies, the National Energy Research Scientific Computing Center, Perimeter Institute, Swiss Federal Institute of Technology Lausanne, the University of Hamburg, the University of Illinois at Urbana–Champaign and the University of Iowa.
Fermilab is a DOE national laboratory supported by the Office of Science.
Recommended reading
A minute with Sajid Ali Syed, computational science and AI research associate
Always looking for a new trail to ride his bike on or an opportunity to catch a live show in town, Sajid Ali Syed is a research associate in the Computational Science and AI Directorate for high-energy physics applications at Fermilab.
DOE invests $13.7 million for research in data reduction for science
The U.S. Department of Energy recently announced $13.7 million in funding for nine research projects that will advance the state of the art in computer science and applied mathematics. One of the recipients of this funding, Fermilab scientist Nhan Tran will lead a project to explore methods for programming custom hardware accelerators for streaming compression.
Coffea speeds up particle physics data analysis
The prodigious amount of data produced at the Large Hadron Collider presents a major challenge for data analysis. Coffea, a Python package developed by Fermilab researchers, speeds up computation and helps scientists work more efficiently. Around a dozen international LHC research groups now use Coffea, which draws on big data techniques used outside physics.