S&TR Features Software for the Exascale Era

The latest issue of LLNL’s Science & Technology Review magazine showcases Computing in the cover story (see abstract below) and Commentary. Open source software plays a prominent role in the initiatives described in the story. The cover art shows an advection simulation powered by open source repos MFEM and GLVis.

• Full issue (and PDF version)
• Commentary: To Exascale and Beyond by Computing associate director Bruce Hendrickson
• Cover story: The Exascale Software Portfolio by Holly Auten and featuring Lori Diachin, Rob Neely, Jeff Hittinger, Ulrike Meier Yang, David Beckingsale, and Tzanio Kolev

As a leader in high-performance computing, Lawrence Livermore wields a large portion of the Department of Energy’s HPC resources to advance national security and foundational science. The Sierra supercomputer supports the National Nuclear Security Administration’s Stockpile Stewardship Program by enabling more accurate, more predictive simulations. This generation of computers is known as heterogeneous, or hybrid, because their architectures combine graphics processing units and central processing units to achieve peak performance well above 100 petaflops. (A petaflop is 10^15 floating-point operations per second.) The next generation’s processing capability—at least an exaflop (10^18 flops)—will be many times greater. HPC software must adjust to these new hardware standards. As the exascale era begins, two major initiatives leverage and expand Livermore’s HPC capabilities, with a spotlight here on software. The Exascale Computing Project, a joint effort between the DOE Office of Science and NNSA, brings together several national laboratories to address many hardware, software, and application challenges inherent in the organizations’ scientific and national security missions. Within the Laboratory, the RADIUSS project aims to benefit scientific applications through a robust software infrastructure.