“MD simulations have significantly increased the pace of developing everything from novel medical therapies to new building materials to better batteries,“ said UC San Diego Chancellor Pradeep K. Khosla. “The San Diego Supercomputer Center at UC San Diego, in collaboration with Intel, is continuing to lead the way to evolve computing and simplify software development so that today’s researchers can work more efficiently to launch tomorrow’s live-saving treatments, sustainable building materials and energy-efficient technologies.”
According to UC San Diego’s Andreas Goetz, assistant research scientist at SDSC and group leader for Data-Driven and High-Performance Computational Chemistry, MD simulations are critical to understanding how molecules, such as proteins and lipids that make up cells or viruses, function on the atomic level, and for predicting how drugs interact with therapeutic targets. Goetz explained that MD time steps are on the order of femtoseconds (one quadrillionth of a second), while relevant physical and biomolecular processes often occur on the microsecond time scale or longer, thus requiring billions of time steps.
“We want to push the boundaries of science to gain insights that are out of reach today, thus, it is critical for us to achieve the highest possible computational performance on any given hardware and to be able to use all the fastest supercomputers in the world, including new exascale machines with multiarchitecture systems including the latest CPUs and GPUs,” said Goetz, who leads the oneAPI Center of Excellence and has been an Amber developer for many years. “The oneAPI Center’s optimizations using oneAPI will modernize Amber’s code for portability making it multiarchitecture-ready and ensuring that MD simulations can take full advantage of the power of upcoming graphics processing units (GPUs).”
An added benefit that oneAPI brings to Amber is the creation of a single source codebase that developers can use to gain flexibility in targeting the best hardware for solutions rather than being confined to a single-vendor proprietary architecture. This is important because the effort to maintain different code bases with vendor-specific programming languages is significant and slows down progress. oneAPI’s open and unified programming model offers a solution to support heterogeneous architectures across multiple vendors.
“We are really excited by this prospect,” said Goetz, “and the lessons learned and best practices for software development will be shared with the broader molecular simulations and computational science community.”
Frank Würthwein, SDSC director and professor of physics at UC San Diego, said that SDSC has a long history of HPC technology collaborations with Intel, and the company’s engineers will work with SDSC and the oneAPI Center on this effort. “We are very happy to establish a close working relationship on open standard programming for high-performance applications like Amber. Our mission at SDSC is to translate innovation into practice and to be at the frontier of research efforts in supercomputing. The oneAPI Center will be important in this regard and foundational to adopt oneAPI for other important high-performance computing applications that our researchers develop,” he said.
“This oneAPI Center of Excellence at the San Diego Supercomputer Center at UC San Diego advances Amber with multiarchitecture readiness via oneAPI. This achievement will benefit thousands of scientists who utilize molecular dynamics simulations to deliver the next breakthroughs in drug discovery and related research,” said Scott Apeland, senior director of Intel Developer Ecosystem Programs.
The San Diego Supercomputer Center is a pioneer and leader in high-performance and data-intensive computing, providing cyberinfrastructure resources, services and expertise to the national research community, academia and industry. Located on the UC San Diego campus, SDSC supports hundreds of multidisciplinary programs spanning a wide variety of domains, from astrophysics and earth sciences to disease research and drug discovery.
oneAPI is an open, unified, cross-architecture programming model for CPUs and accelerator architectures (GPUs, FPGAs, and others). Based on standards, the programming model simplifies software development and delivers uncompromised performance for accelerated compute without proprietary lock-in, while enabling the integration of existing code. With oneAPI, developers can choose the best architecture for the specific problem they are trying to solve without needing to rewrite software for the next architecture and platform.