Video available here: https://youtu.be/-scLUoHmFo4?t=3420 Targeting Real chemical accuracy at the EXascale Quantum mechanical simulations play an important role in scientific and industrial applications, where understanding at the atomistic and electronic level is needed. In this context, density functional theory is generally the method of choice thanks to its relatively low computational cost and ability to appropriately describe the electronic properties of a variety of systems. Despite the many successes, there are however materials or classes of materials properties (e.g. magnetic systems, photo-induced processes, non-covalent interactions), where the mean-field picture becomes insufficient, and the fully correlated nature of the system must be considered. The computational cost of high-accuracy quantum mechanical approaches is orders of magnitudes larger than the cost of density functional theory, and their implementation must be reconsidered in order to take advantage of the architecture of modern and future supercomputers. The TREX European Center of Excellence (CoE) will start in October. It aims at developing and promoting an open-source, high-performance software platform in the domain of high-accuracy (stochastic) quantum chemical simulations, designed for the upcoming Exascale architectures. Here, we will present the main aspects of the TREX CoE and, in particular, describe our strategy to design a high-performance library for quantum Monte Carlo simulations
Video available here: https://youtu.be/-scLUoHmFo4?t=3420 Targeting Real chemical accuracy at the EXascale Quantum mechanical simulations play an important role in scientific and industrial applications, where understanding at the atomistic and electronic level is needed. In this context, density functional theory is generally the method of choice thanks to its relatively low computational cost and ability to appropriately describe the electronic properties of a variety of systems. Despite the many successes, there are however materials or classes of materials properties (e.g. magnetic systems, photo-induced processes, non-covalent interactions), where the mean-field picture becomes insufficient, and the fully correlated nature of the system must be considered. The computational cost of high-accuracy quantum mechanical approaches is orders of magnitudes larger than the cost of density functional theory, and their implementation must be reconsidered in order to take advantage of the architecture of modern and future supercomputers. The TREX European Center of Excellence (CoE) will start in October. It aims at developing and promoting an open-source, high-performance software platform in the domain of high-accuracy (stochastic) quantum chemical simulations, designed for the upcoming Exascale architectures. Here, we will present the main aspects of the TREX CoE and, in particular, describe our strategy to design a high-performance library for quantum Monte Carlo simulations
Video available here: https://youtu.be/-scLUoHmFo4?t=3420 Targeting Real chemical accuracy at the EXascale Quantum mechanical simulations play an important role in scientific and industrial applications, where understanding at the atomistic and electronic level is needed. In this context, density functional theory is generally the method of choice thanks to its relatively low computational cost and ability to appropriately describe the electronic properties of a variety of systems. Despite the many successes, there are however materials or classes of materials properties (e.g. magnetic systems, photo-induced processes, non-covalent interactions), where the mean-field picture becomes insufficient, and the fully correlated nature of the system must be considered. The computational cost of high-accuracy quantum mechanical approaches is orders of magnitudes larger than the cost of density functional theory, and their implementation must be reconsidered in order to take advantage of the architecture of modern and future supercomputers. The TREX European Center of Excellence (CoE) will start in October. It aims at developing and promoting an open-source, high-performance software platform in the domain of high-accuracy (stochastic) quantum chemical simulations, designed for the upcoming Exascale architectures. Here, we will present the main aspects of the TREX CoE and, in particular, describe our strategy to design a high-performance library for quantum Monte Carlo simulations
TREX : an innovative view of HPC usage applied to Quantum Monte Carlo simulations The TREX[1] European Center of Excellence focuses on high accuracy quantum mechanical methods, essential in many different fields of application such as new material design or photochemistry. Among these methods, Quantum Monte Carlo (QMC) approaches are particularly well adapted to exascale architectures. Our ambition is to the help the community take advantage of exascale machines through the use of our HPC software. We will review in the presentation progress along the three following axes: * TREXIO[2]: A common I/O library and file format for easily exchanging data between applications, facilitating high-throughput computing workflows, * QMCkl[3]: A library of computational kernels specific to QMC applications written together by QMC and HPC experts, taking advantage of both CPUs and GPUs, * An integrated workflow including performance analysis (MAQAO[4]) and numerical accuracy measurements (Verificarlo[5]) to be used for the development of QMC kernels and more generally for improving the applications. In particular, we plan to identify the best performance usage of QMCkl and also to adjust the performance with numerical precision requirements. ------------------------------------ [1] https://trex-coe.eu [2] https://github.com/trex-coe/trexio [3] https://trex-coe.github.io/qmckl [4] https://www.maqao.org [5] https://github.com/verificarlo ------------------------------------
There are three different Quantum Monte Carlo (QMC) codes in the Targeting REal chemical accuracy at the eXascale (TREX) Center of Excellence (CoE), TurboRVB, Champ, and QMC=Chem, each with its own strengths and weaknesses. Instead of optimizing the three codes independently for exascale architectures, or re-writing a new monolithic code, our strategy is instead to design a new library, Quantum Monte Carlo kernel library (QMCkl), containing the state-of-the-art expertise in the implementation of specific kernels present in each of the three codes. The functions available in this library will allow all the codes to benefit from the optimal implementation of the major kernels of QMC. The three main objectives driving the development of QMCkl are performance, productivity and portability.
The quest for chemical accuracy in predicting molecular properties and reactivity remains a central challenge in computational chemistry. While Density Functional Theory (DFT) offers a favorable cost-to-accuracy ratio, its inherent approximations limit its ability to consistently achieve sub-kJ/mol accuracy, particularly for complex systems. Conversely, high-level wave function theory (WFT) methods, often termed "post-Hartree-Fock" (post-HF) methods, deliver unparalleled accuracy but are prohibitively expensive for systems beyond a few dozen atoms. This paper explores the synergistic potential of hybrid DFT/post-HF schemes in the context of emerging exascale computing architectures. We delineate how combining DFT's efficiency for electron correlation and the rigorous treatment of dynamic and static correlation by post-HF methods can overcome individual limitations. Exascale computing, with its unprecedented power and massive parallelism, provides a unique opportunity to accelerate these sophisticated hybrid methods. We discuss algorithmic challenges, software engineering requirements, and novel parallelization strategies to harness exascale capabilities effectively. By detailing specific hybrid schemes, such as double-hybrid functionals and explicitly correlated methods, we demonstrate their promise for pushing the boundaries of chemical accuracy for previously inaccessible systems. The ultimate goal is to enable predictive simulations that drive discovery in materials science, drug design, and catalysis, making chemical accuracy a routine harvest rather than an exceptional yield.
The era of exascale computing presents both exciting opportunities and unique challenges for quantum mechanical simulations. While the transition from petaflops to exascale computing has been marked by a steady increase in computational power, the shift towards heterogeneous architectures, particularly the dominant role of graphical processing units (GPUs), demands a fundamental shift in software development strategies. This review examines the changing landscape of hardware and software for exascale computing, highlighting the limitations of traditional algorithms and software implementations in light of the increasing use of heterogeneous architectures in high-end systems. We discuss the challenges of adapting quantum chemistry software to these new architectures, including the fragmentation of the software stack, the need for more efficient algorithms (including reduced precision versions) tailored for GPUs, and the importance of developing standardized libraries and programming models.
The era of exascale computing presents both exciting opportunities and unique challenges for quantum mechanical simulations. While the transition from petaflops to exascale computing has been marked by a steady increase in computational power, the shift towards heterogeneous architectures, particularly the dominant role of graphical processing units (GPUs), demands a fundamental shift in software development strategies. This review examines the changing landscape of hardware and software for exascale computing, highlighting the limitations of traditional algorithms and software implementations in light of the increasing use of heterogeneous architectures in high-end systems. We discuss the challenges of adapting quantum chemistry software to these new architectures, including the fragmentation of the software stack, the need for more efficient algorithms (including reduced precision versions) tailored for GPUs, and the importance of developing standardized libraries and programming models.