Lattice Field Theory (LFT) provides tools for studying the fundamental forces of nature using numerical simulations. The traditional realm of application of LFT has been Quantum Chromodynamics (QCD), the theory describing the strong nuclear force within the Standard Model (SM) of particle physics. These calculations now include electromagnetic effects and achieve sub-percent accuracy. Other applications span a wide range of topics, from theories beyond the Standard Model, to low-dimensional strongly coupled fermionic models, to new cosmological paradigms. At the core of this scientific endeavour lies the ability to perform sophisticated and demanding numerical simulations. The exascale era of High-Performance Computing therefore is going to be a time of great opportunity for fundamental particle physics research.
The UK research programme in LFT, and its impact, can be expanded in a transformative way with the advent of pre-exascale and exascale systems, but only if key challenges are addressed. As the number of floating-point operations per second increases, communication between computing nodes is lagging behind, and this imbalance will severely affect future LFT simulations across the board. The scientific impact of all lattice studies is dictated by the ability to tame systematic errors by simulating finer lattice spacings, larger physical volumes, and smaller fermion masses. These multiscale problems are universal to LFT and will need to be addressed, while guaranteeing portability across heterogeneous architectures without committing to a specific vendor or technology solution.
The goal of the EXA-LAT project is to develop a common set of best practices, KPIs and figures of merit that can be used by the whole LFT community in the near future to inform the design and procurement of future systems.
The physics objectives can be divided into three main categories that map closely to the current experimental programme in particle physics:
- To understand the fundamental symmetries determining the structure of the universe and search for new interactions Beyond the Standard Model (BSM): discovering the basic forces of nature and understanding their dynamics are at the forefront of physics research.
- To understand the phases of strongly interacting matter and the properties of the quark-gluon plasma (QGP): This is one of the key missions of the nuclear physics program in Europe and internationally.
- To describe the structure and interactions of hadrons and light nuclei from QCD.
New simulations using large volumes, with fine lattice spacings and small masses will provide transformative input for all these programmes.
Working Group Objectives:
- Developing a roadmap for a broad community to be able to exploit exascale facilities by the mid-2020s, tying together requirements for code, algorithmic and hardware developments to physics targets.
- Benchmarking and identification of bottlenecks on different systems. The objective here is to define a set of common benchmarks, KPIs and Figures of Merit that will allow the working group to quantify progress towards the exascale in the forthcoming years. Having identified the bottlenecks, the working group will start designing and implementing optimised kernels.
- Developing new algorithms together with mathematicians and computer scientists, to break new ground in domain-specific problems: signal-to-noise ratio, multigrid inverters and solving instability challenges of the current algorithms when scaling to large volumes. Explore new directions that use AI for LFT and vice versa, i.e. that use LFT methods to improve on AI problems.
