Intel oneAPI empowers GROMACS 2022 to promote the development of open source drugs

GROMACS is accelerated by Intel oneAPI open programming and multi-architecture tools and runs on Intel Xe architecture-based GPUs, demonstrating outstanding performance.

Intel is committed to fostering an open ecosystem, including making technical contributions to many open source projects that have a direct impact on the real world. One example is GROMACS , a molecular dynamics software package used to simulate proteins, lipids, and nucleic acids for designing new drugs. The recently released GROMACS 2022 was developed using SYCL and oneAPI, and it shows excellent performance running on multiple architectures, including GPUs based on Intel Xe architecture.

“GROMACS is one of the most widely used open source molecular dynamics applications in the world. The reason is simple: simulations run with it help us better understand the dynamics of molecules ranging from proteins in the body to galaxies in the universe,” said Roland Schulz, parallel software engineer at Intel. “Remarkably, our collaboration with GROMACS, developed and optimized using oneAPI, has allowed Intel to participate in major advances in drug discovery and development, and has expanded the open development of GROMACS across multiple computing architectures. All of this has been achieved while collaborating with the open source community that we hold so dear.”

GROMACS molecular dynamics simulations are powered by oneAPI, and together with international distributed computing projects such as Folding@home, have contributed to identifying solutions for key drugs for diseases such as breast cancer, novel coronavirus pneumonia, and type 2 diabetes. In modern drug development, molecular dynamics simulations have been widely and successfully used. These simulations provide researchers with the structural information of biomacromolecules needed to understand the relationship between structure and function, which is used to guide the drug development and design process. The application of computational tools like GROMACS in drug development helps researchers design and evaluate new drugs more efficiently while saving resources.

The GROMACS R&D team at Stockholm University and the Royal Institute of Technology, led by Professor of Biophysics Erik Lindahl, leads the development of the GROMACS molecular dynamics toolkit, one of the most widely used high-performance computing applications in the world. Molecular dynamics is computationally centric and iterative, making it one of the most time-consuming high-performance computing applications. Billions of calculations involve millions of lines of code.

oneAPI, as an open, unified programming model for CPUs and accelerators, and supporting computer architectures from multiple vendors, helped Professor Lindahl and his team expand GROMACS support for heterogeneous hardware. This is due to the use of cross-architecture, cross-vendor open standards to improve productivity. Based on these standards, oneAPI programming simplifies the software development process, providing accelerated computing performance without the need for a specific programming language or vendor, while allowing the integration of existing code, including OpenMP.

As part of the oneAPI optimization work, Lindahl's team migrated the CUDA code in GROMACS that could only run on Nvidia hardware to SYCL by using the Intel® DPC++ Compatibility Tool (part of the Intel® oneAPI Base Toolkit), which typically automates 90 to 95% of the code migration work1,2. This allowed his team to create a new, independently portable cross-architecture code base. This greatly simplified development work and provided flexibility for deployment in multi-architecture environments.

“With full support for SYCL and oneAPI in GROMACS 2022, we have extended GROMACS to run on new types of hardware,” said Lindahl. “Through Intel® DevCloud, we have already run production simulations on current Intel Xe architecture-based GPUs, as well as on Ponte Vecchio, our upcoming Intel Xe architecture-based GPU development platform. The performance results achieved in this phase are impressive and demonstrate the power of Intel’s hardware and software collaboration. Altogether, these optimizations enable hardware diversity, deliver high-end performance, and drive competition and innovation, allowing us to conduct science faster and reduce costs for downstream industries.”

GROMACS achieves accelerated computing through optimization using Intel oneAPI cross-architecture tools, such as the oneAPI DPC++/C++ compiler, oneAPI library, and high-performance computing analysis and cluster tools. The oneAPI tools are available on Intel® DevCloud, a free environment for developing and testing code on various Intel architectures such as CPU, GPU, and FPGA.