About this Research Topic
Exascale computing is defined as the capability of calculating 1018 floating point operations per second (exaflops), which is five orders of magnitude higher than the capability of a state-of-the-art workstation. Summit and Fugaku supercomputers already marked the onset of the exascale era by reaching more than one exaflops with reduced precision in 2020. Meanwhile, several new supercomputers that can sustainably reach more than one exaflops in double precision are expected to be operational within two years. Clearly, this provides unprecedented opportunities for computational chemists such as longer simulation times in molecular dynamics, faster time to solution in electronic structure calculations, and more realistic environments in multi-scale modelling. Moreover, data-driven projects coupling machine learning and simulations can thrive in this environment since these machines are being optimized for AI workloads. However, utilizing these new supercomputers efficiently is challenging due to various architectural differences of these machines compared to typical workstations used by computational chemists.
In this Research Topic we would like to identify the opportunities and challenges for computational chemists in the exascale era and report prospective solutions that can help researchers utilize these resources efficiently. High-performance computing centers around the world funded by national or international institutions have programs to develop and optimize application software for their systems and they work closely with many different application developers. The rise of graphical processing units (GPUs) for scientific computing and breakthroughs in machine learning were two paradigm changes that requires a special focus for the exascale era. Exascale Computing Project is a nationwide effort within the USA, that support application development and hardware software integration for this special era. Similar projects have been launched in Europe and Asia to overcome the challenges of this transition. This research topic will provide a medium for these researchers to share their solutions to a broader community that can make use of their findings. Particularly, we will focus on methodological and algorithmic improvements, optimizations of application software, use of multi-precision methods and complex workflows that couples simulation and machine learning. We welcome Original Research, Review, Mini Review and Perspective articles on themes including, but not limited to:
• Novel computational chemistry methods and implementations with an analysis on weak and strong scaling, memory bandwidth, and precision characteristics,
• Workflows that couples simulations and machine learning,
• Developments in molecular dynamics and quantum chemistry codes,
• Single and multi-GPU optimizations with different programming models,
• Profiling and optimization for distributed memory,
• Reduced and multi-precision algorithms,
• Portability and containerization for application software.
In this Research Topic we would like to identify the opportunities and challenges for computational chemists in the exascale era and report prospective solutions that can help researchers utilize these resources efficiently. High-performance computing centers around the world funded by national or international institutions have programs to develop and optimize application software for their systems and they work closely with many different application developers. The rise of graphical processing units (GPUs) for scientific computing and breakthroughs in machine learning were two paradigm changes that requires a special focus for the exascale era. Exascale Computing Project is a nationwide effort within the USA, that support application development and hardware software integration for this special era. Similar projects have been launched in Europe and Asia to overcome the challenges of this transition. This research topic will provide a medium for these researchers to share their solutions to a broader community that can make use of their findings. Particularly, we will focus on methodological and algorithmic improvements, optimizations of application software, use of multi-precision methods and complex workflows that couples simulation and machine learning. We welcome Original Research, Review, Mini Review and Perspective articles on themes including, but not limited to:
• Novel computational chemistry methods and implementations with an analysis on weak and strong scaling, memory bandwidth, and precision characteristics,
• Workflows that couples simulations and machine learning,
• Developments in molecular dynamics and quantum chemistry codes,
• Single and multi-GPU optimizations with different programming models,
• Profiling and optimization for distributed memory,
• Reduced and multi-precision algorithms,
• Portability and containerization for application software.
Keywords: High-performance computing, GPU optimization, Exascale era
Important Note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.
