Parallel Supercomputers Research Articles

Data-driven sensitivity analysis in surface structure determination using total-reflection high-energy positron diffraction (TRHEPD)

The present article proposes a data analysis method for experimentally-derived measurements, which consists of an auto-optimization procedure and a sensitivity analysis. The method was applied to the results of a total-reflection high-energy positron diffraction (TRHEPD) experiment, a novel technique of determining surface structures or the position of the atoms near the material surface. This method solves numerically the partial differential equation in the fully-dynamical quantum diffraction theory with many trial surface structures. In the sensitivity analysis, we focused on the experimental uncertainties and the variation over individual fitting parameters, which was analyzed by solving the eigenvalue problem of the variance-covariance matrix. A modern massively parallel supercomputer was used to complete the analysis within a moderate computational time. The sensitivity analysis provides a basis for the choice of variables in the data analysis for practical reliability. The effectiveness of the present analysis method was demonstrated in the structure determination of a Si4O5N3 / 6H-SiC(0001)-(3×3) R30∘ surface. Furthermore, this analysis method is applicable to many experiments other than TRHEPD.

Simulations of future particle accelerators: issues and mitigations

The ever increasing demands placed upon machine performance have resulted in the need for more comprehensive particle accelerator modeling. Computer simulations are key to the success of particle accelerators. Many aspects of particle accelerators rely on computer modeling at some point, sometimes requiring complex simulation tools and massively parallel supercomputing. Examples include the modeling of beams at extreme intensities and densities (toward the quantum degeneracy limit), and with ultra-fine control (down to the level of individual particles). In the future, adaptively tuned models might also be relied upon to provide beam measurements beyond the resolution of existing diagnostics. Much time and effort has been put into creating accelerator software tools, some of which are highly successful. However, there are also shortcomings such as the general inability of existing software to be easily modified to meet changing simulation needs. In this paper possible mitigating strategies are discussed for issues faced by the accelerator community as it endeavors to produce better and more comprehensive modeling tools. This includes lack of coordination between code developers, lack of standards to make codes portable and/or reusable, lack of documentation, among others.

A robust and scalable unfitted adaptive finite element framework for nonlinear solid mechanics

In this work, we bridge standard Adaptive Mesh Refinement and coarsening (AMR) on scalable octree background meshes and robust unfitted Finite Element (FE) formulations for the automatic and efficient solution of large-scale nonlinear solid mechanics problems posed on complex geometries, as an alternative to standard body-fitted formulations, unstructured mesh generation and graph partitioning strategies. We pay special attention to those aspects requiring a specialized treatment in the extension of the unfitted h-adaptive Aggregated Finite Element Method (h-AgFEM) on parallel tree-based adaptive meshes, recently developed for linear scalar elliptic problems, to handle nonlinear problems in solid mechanics. In order to accurately and efficiently capture localized phenomena that frequently occur in nonlinear solid mechanics problems, we perform pseudo time-stepping in combination with h-adaptive dynamic mesh refinement and re-balancing driven by a-posteriori error estimators. The method is implemented considering both irreducible and mixed (u/p) formulations and thus it is able to robustly face problems involving incompressible materials. In the numerical experiments, both formulations are used to model the inelastic behavior of a wide range of compressible and incompressible materials. First, a selected set of benchmarks is reproduced as a verification step. Second, a set of experiments is presented with problems involving complex geometries. Among them, we model a cantilever beam problem with spherical hollows distributed in a simple cubic (SC) array. This test involves a discrete domain with up to 11.7M Degrees Of Freedom (DOFs) solved in less than two hours on 3072 cores of a parallel supercomputer.

A conservative multirate explicit time integration method for computation of compressible flows

In the context of high fidelity simulation of compressible flows (LES and DNS) at extreme scale (small time steps) on massively parallel supercomputers, explicit time integration methods are widely used since they both have a good computational cost trade-off at small time scales and require a small number of parallel communications, thus having little impact on the parallel strategy compared to their implicit counterpart. However, the synchronous nature of the time integration of the governing equations can be a severe constraint when the stability condition is applied globally since the time scales are related to the flow properties and the mesh resolution, which may show strong variations throughout the computational domain. We propose to further improve the efficiency (CPU wall-time reduction) of explicit Runge–Kutta methods by developing a multirate explicit time integration method, by means of flux interpolation at the boundary between cells evolving with different time-steps, which enforces the conservation properties. In terms of computational efficiency, the presented multirate time integration method is easy to implement in pre-existing Eulerian compressible Navier–Stokes codes, requires less additional memory storage, and provides a considerable speed-up while being robust and preserving the order of accuracy of the legacy explicit Runge–Kutta time integration method. The multirate time integration method is implemented in the massively parallel finite volume and high-order (spectral difference) IC3 code (Bodart et al., 2016) (fork of the solver CharLESX), but it can also be applied to any flux-based spatial method such as discontinuous Galerkin or others. For a targeted y+=0.2 on the developed turbulent channel flow test case at Reτ=392; a 2.48 effective speedup is obtained versus an expected theoretical speedup of 2.53.

TinyMD: Mapping molecular dynamics simulations to heterogeneous hardware using partial evaluation

This paper investigates the suitability of the AnyDSL partial evaluation framework to implement tinyMD: an efficient, scalable, and portable simulation of pairwise interactions among particles. We compare tinyMD with the miniMD proxy application that scales very well on parallel supercomputers. We discuss the differences between both implementations and contrast miniMD’s performance for single-node CPU and GPU targets, as well as its scalability on SuperMUC-NG and Piz Daint supercomputers. Additionally, we demonstrate tinyMD’s flexibility by coupling it with the waLBerla multi-physics framework. This allow us to execute tinyMD simulations using the load-balancing mechanism implemented in waLBerla.

Turbulent flow characteristics around a non-submerged rectangular obstacle on the side of an open channel

The three-dimensional flow structure and turbulence characteristics around a non-submerged rectangular obstacle in an open channel are explored using numerical simulation. In particular, a low length-to-depth ratio condition, shown to be associated with three-dimensional flow features in our previous study, is considered. To sufficiently resolve all the important details of the three-dimensional turbulent flow around and in the entire wake of an obstacle, high-resolution large-eddy simulation (LES) employing is carried out on a parallel supercomputer. The LES results were compared with the measurements and analyzed to examine the horseshoe vortex structure, free-surface vortex structure, recirculation zones, vortex shedding process, turbulence characteristics, and wall shear stress distribution around an obstacle. The results provide important insights into the complete three-dimensional flow structure and wall shear stress patterns around the obstacle.

Transition region adaptive conduction (TRAC) in multidimensional magnetohydrodynamic simulations

Context. In solar physics, a severe numerical challenge for modern simulations is properly representing a transition region between the million-degree hot corona and a much cooler plasma of about 10 000 K (e.g., the upper chromosphere or a prominence). In previous 1D hydrodynamic simulations, the transition region adaptive conduction (TRAC) method has been proven to capture aspects better that are related to mass evaporation and energy exchange. Aims. We aim to extend this method to fully multidimensional magnetohydrodynamic (MHD) settings, as required for any realistic application in the solar atmosphere. Because modern MHD simulation tools efficiently exploit parallel supercomputers and can handle automated grid refinement, we design strategies for any-dimensional block grid-adaptive MHD simulations. Methods. We propose two different strategies and demonstrate their working with our open-source MPI-AMRVAC code. We benchmark both strategies on 2D prominence formation based on the evaporation–condensation scenario, where chromospheric plasma is evaporated through the transition region and then is collected and ultimately condenses in the corona. Results. A field-line-based TRACL method and a block-based TRACB method are introduced and compared in block grid-adaptive 2D MHD simulations. Both methods yield similar results and are shown to satisfactorily correct the underestimated chromospheric evaporation, which comes from a poor spatial resolution in the transition region. Conclusions. Because fully resolving the transition region in multidimensional MHD settings is virtually impossible, TRACB or TRACL methods will be needed in any 2D or 3D simulations involving transition region physics.

Research on Computer System Implementation of ISO14443B Protocol Communication Data Acquisition based on FPGA

As a programmable logic device, FPGA can be used as a bridge for logic adaptation and timing matching between chip components. For example, in a data acquisition system, data of ten channels need to be collected at the same time. In this case, ten identical data acquisition modules can be instantiated in the FPGA to connect ten external data sensor chips respectively by taking advantage of the parallelism of FPGA, so as to realize the logical adaptation of chip devices. In certain applications such as audio processing, video processing, digital communications, encryption, decryption, etc., as a result of the FPGA computing can parallelism, and FPGA is algorithm Yu into hardware, namely a logic gate and a register, so its execution efficiency usually than machine code and assembly language of single thread special micro processing, the need to replace DSP parallel supercomputing occasions and microprocessors.

Comments to Jean-Claude Burgelman's article Politics and Open Science: How the European Open Science Cloud Became Reality (the Untold Story)

Is the evolution of how science is done primarily heuristic or driven by policy makers? Reading the intriguing story of the inception of the European Open Science Cloud (EOSC) by J-C Burgelman one could assume the latter. But the story also reveals that the policy makers' shaping of concepts and planning of funding models is a mere reaction to the turns of the broad stream of research. Looking from the science side the trend towards data intensive science started for sure already in the 90's with advancements in particularly sensor technologies rapidly increasing the amount of data collected. The development of Charge-coupled Devices (CCDs) and its impact on astronomy is the often-cited example but it was a trend building up all over experimental research taking benefit of developments within fast electronics and communication, or leaps in efficiency when shifting experimental technology, e.g. going from chromatography methods to fluoresces based methods for genome sequencing. On top of it the introduction of parallel supercomputers almost completely built out of commercial off-the-shelf components enabled the possibility to actually analyse and produce data in an unprecedented cost-efficient way.The resulting obvious need to combine and interoperate various networked resources was met by computer science and maybe the most successful work was concluded in the “The GRID, Blueprint for a New Computing Infrastructure”①. This result of work during the 90's by Foster and Kesselman became immensely popular even in industry. Though, on the other side of the IT-bubble industry took another route. Introduction of Service Oriented Architecture (SOA), Web-services, and cloud computing became the industry standard implementation of the ubiquitous computing promise in “The GRID”. Soon it also became part of the researchers' toolbox. Meaning that more complex workflows could be realised in connecting remote experiments, data resources, computing facilities, and whatever needed to reach the objectives of and advanced the research enterprise.The European Commission supported a large number of Grid-projects during the first decade of the millennium and the technology and concepts were particularly adopted by the high-energy physics community which still use them for analysing and simulating data related to the Large Hadron Collider (LHC) through a global network, or grid of sites.It deserves a more careful investigation but it is my feeling that particularly the life-science community was leading in taking another route and jumped on the train of SOA, Web-services, and cloud computing when it started to move with some speed somewhere after 2005. The inherit need of the specific research and the smart utilisation of the available technologies have resulted in a very well organised community when it comes to gathering of common open core data resources; sharing results that are findable, accessible, interoperable, and reusable; and very much of everything else that we are seeking with EOSC. In this aspect, a life-science, or definitely a bioinformatics EOSC implementation is already in place and also heavily used by the pharmaceutical industry which can access many of the core data resources at the same terms and conditions as publicly funded research.What is the point in looking backwards for some more than a decade old advances? Just to show that where we are now is not a new place. Yes, we experience another turn in how research is done and we are now looking for a societal response to that change. How can the society at large support, but also take optimal benefit of where research is right now? EOSC is, if we succeed, part of the answer, but we need to make sure that it can stay creative and innovative in enabling opportunities for research and society at large to thrive together.High-energy physics and life-sciences are mentioned as areas driving the development of concepts for enabling of research, but examples can be found elsewhere. More important is what research areas and research needs will show the direction for the next leap, maybe humanities or environmental sciences? I would put my bet there somewhere but the important message is that inspiration and innovation will, and must come from the research needs. Only then can growth minded policy makers, in their best moments, accelerate the development by introducing new frameworks and steering of funding, that will ultimately benefit research and the whole society.

Parallel Code Execution as a Tool for Enhancement of the Sustainable Design of Foundation Structures

Civil engineering structures are always in interaction with the underlying parts of the Earth. This form of interaction results in deformations and stresses that affect the service life of structures. Long and predictable service life is one of important aspects of sustainable design. Thus, good knowledge of the interaction effects is an essential part of sustainable design. However, to obtain this information, the use of complex numerical models is often necessary. In many cases, the creation and analysis of such complex models are not possible with the tools usually available in civil engineering practice. Technically, the necessary software and computer hardware exist, but their use for such tasks is still very infrequent and includes many challenges. The main aim of this article was thus to propose an approach of numerical analysis that utilizes parallel supercomputers and software based on the finite element method. The paper concentrated on the feasibility of the solution and on calculation times, because these aspects are usually the main reasons why engineers in practice tend to reject these approaches. The approach was demonstrated on a model case that was compatible with actual in situ experiments executed by the author’s team, and thus the validity of the computed results is verifiable. Limitations of the proposed approach are also discussed.

Application-specific feature selection and clustering approach with HPC system profiling data

Exascale computing, the next-generation computing environment, is expected to be applied to scientific and engineering applications. Accordingly, high-performance computing (HPC) technology is also being developed to improve the performance and high-speed parallelism of many-core processors. Previous researches on improving HPC performance have developed in the form of improving the overall system performance by analyzing the state of the system occurring in the range of the knowledge of expert. However, performance events occurring in a processor in a many-core environment have a large number of indicators, and it is difficult to analyze the correlation between them. In this paper, we propose an application-specific feature selection and clustering approach with HPC system profiling data. The proposed approach performs PCA-based feature selections for efficient performance analysis methods. In addition, the application-specific characteristics from profiling data can be analyzed by unsupervised learning. In our experiments, we evaluated highly parallel supercomputers with NAS parallel benchmark and were able to cluster applications efficiently.

Supercomputer-Based Ensemble Docking Drug Discovery Pipeline with Application to Covid-19.

We present a supercomputer-driven pipeline for in silico drug discovery using enhanced sampling molecular dynamics (MD) and ensemble docking. Ensemble docking makes use of MD results by docking compound databases into representative protein binding-site conformations, thus taking into account the dynamic properties of the binding sites. We also describe preliminary results obtained for 24 systems involving eight proteins of the proteome of SARS-CoV-2. The MD involves temperature replica exchange enhanced sampling, making use of massively parallel supercomputing to quickly sample the configurational space of protein drug targets. Using the Summit supercomputer at the Oak Ridge National Laboratory, more than 1 ms of enhanced sampling MD can be generated per day. We have ensemble docked repurposing databases to 10 configurations of each of the 24 SARS-CoV-2 systems using AutoDock Vina. Comparison to experiment demonstrates remarkably high hit rates for the top scoring tranches of compounds identified by our ensemble approach. We also demonstrate that, using Autodock-GPU on Summit, it is possible to perform exhaustive docking of one billion compounds in under 24 h. Finally, we discuss preliminary results and planned improvements to the pipeline, including the use of quantum mechanical (QM), machine learning, and artificial intelligence (AI) methods to cluster MD trajectories and rescore docking poses.

Integrating state of the art compute, communication, and autotuning strategies to multiply the performance of ab initio molecular dynamics on massively parallel multi-core supercomputers

The development in today’s supercomputer hardware is that the compute power of the individual nodes grows much faster than the speed of their interconnects. To benefit from this evolution in computer hardware, the challenge in modernization of simulation software is to increase the computational load on the nodes and to reduce simultaneously the inter-node communication. Here, we demonstrate the implementation of such a strategy for plane-wave based electronic structure methods and ab initio molecular dynamics (AIMD) simulations. Our focus is on ultra-soft pseudopotentials (USPP), since they allow to shift workload from fast Fourier transforms (FFTs) to highly node-efficient matrix–matrix multiplications. For communication intensive routines, as the multiple distributed 3-d FFTs of the electronic states and the distributed matrix–matrix multiplications related to the β-projectors of the pseudopotentials, parallel MPI+OpenMP algorithms are revised to make use of overlapping computation and communication. The necessary partitioning of the workload is optimized by auto-tuning algorithms. In addition, the largest global MPI_Allreduce operation is replaced by highly tuned node-local parallelized operations using MPI shared-memory windows to avoid inter-node communication. A batched algorithm for the multiple 3-d FFTs improves the throughput of the MPI_Alltoall communication and, thus, the scalability of the implementation, both for USPP and for frequently used norm-conserving pseudopotentials. The new algorithms have been implemented in the AIMD program CPMD (www.cpmd.org). The enhanced performance and scalability of the code is demonstrated on simulations of liquid water with up to 2048 molecules. It is shown that 100ps simulations with many hundred water molecules can be done now routinely within a few days on a moderate number of nodes.

Simflowny 3: An upgraded platform for scientific modeling and simulation

Simflowny is an open platform which automatically generates efficient parallel code of scientific dynamical models for different simulation frameworks. Here we present major upgrades on this software to support simultaneously a quite generic family of partial differential equations. These equations can be discretized using: (i) standard finite-difference for systems with derivatives up to any order, (ii) High-Resolution-Shock-Capturing methods to deal with shocks and discontinuities of balance law equations, and (iii) particle-based methods. We have improved the adaptive-mesh-refinement algorithms to preserve the convergence order of the numerical methods, which is a requirement for improving scalability. Finally, we have also extended our graphical user interface (GUI) to accommodate these and future families of equations. This paper summarizes the formal representation and implementation of these new families, providing several validation results. Program summaryProgram Title: SimflownyCPC Library link to program files:https://doi.org/10.17632/g9mcw8s64f.2Licensing provisions: Apache License, 2.0Programming language: Java, C++ and JavaScriptJournal Reference of previous version: Comput. Phys. Comm. 184 (2013) 2321–2331, Comput. Phys. Comm. 229 (2018), 170–181Does the new version supersede the previous version?: YesReasons for the new version: Additional featuresSummary of revisions: Expanded support for Partial Differential Equations, meshless particles and advanced Adaptive Mesh Refinement.Nature of problem: Simflowny generates numerical simulation code for a wide range of models.Solution method: Any discretization scheme based on either Finite Volume Methods, Finite Difference Methods, or meshless methods for Partial Differential Equations.Additional comments: Simflowny runs in any computer with Docker [1]. Installation details can be checked in the documentation of Simflowny [2]. It can also be compiled from scratch on any Linux system, provided dependences are properly installed as indicated in the documentation. The generated code runs on any Linux platform ranging from personal workstations to clusters and parallel supercomputers. The software architecture is easily extensible for future additional model families and simulation frameworks. Full documentation is available in the wiki home of the Simflowny project [2].

A New Rusanov-Type Solver with a Local Linear Solution Reconstruction for Numerical Modeling of White Dwarf Mergers by Means Massive Parallel Supercomputers

The results of numerical modeling of white dwarf mergers on massive parallel supercomputers using a AVX-512 technique are presented. A hydrodynamic model of white dwarfs closed by a star equation of state and supplemented by a Poisson equation for the gravitational potential is constructed. This paper presents a modification based on a local linear reconstruction of the solution of the Rusanov scheme for the hydrodynamic equations. This reconstruction makes it possible to considerably decrease the numerical dissipation of the scheme for weak shock waves without any external piecewise polynomial reconstruction. The scheme is efficient for unstructured grids, when it is difficult to construct a piecewise polynomial solution, and also in parallel implementations of structured nested or adaptive grids, when the costs of interprocess interactions increase significantly. As input data, piecewise constant values of the physical variables in the left and right cells of a discontinuity are used. The smoothness of the solution is measured by the discrepancy between the maximum left and right eigenvalues. This discrepancy is used for a local piecewise polynomial reconstruction in the left and right cells. Then the solutions are integrated along the characteristics taking into account the piecewise linear representation of the physical variables. A performance of 234 gigaflops and 33-fold speedup are obtained on two Intel Skylake processors on the cluster NKS-1P of the Siberian Supercomputer Center ICM & MG SB RAS.

Scalable molecular dynamics on CPU and GPU architectures with NAMD.

NAMDis a molecular dynamics program designed for high-performance simulations of very large biological objects on CPU- and GPU-based architectures. NAMD offers scalable performance on petascale parallel supercomputers consisting of hundreds of thousands of cores, as well as on inexpensive commodity clusters commonly found in academic environments. It is written in C++ and leans on Charm++ parallel objects for optimal performance on low-latency architectures. NAMD is a versatile, multipurpose code that gathers state-of-the-art algorithms to carry out simulations in apt thermodynamic ensembles, using the widely popular CHARMM, AMBER, OPLS, and GROMOS biomolecular force fields. Here, we review the main features of NAMD that allow both equilibrium and enhanced-sampling molecular dynamics simulations with numerical efficiency. We describe the underlying concepts utilized by NAMD and their implementation, most notably for handling long-range electrostatics; controlling the temperature, pressure, and pH; applying external potentials on tailored grids; leveraging massively parallel resources in multiple-copy simulations; and hybrid quantum-mechanical/molecular-mechanical descriptions. We detail the variety of options offered by NAMD for enhanced-sampling simulations aimed at determining free-energy differences of either alchemical or geometrical transformations and outline their applicability to specific problems. Last, we discuss the roadmap for the development of NAMD and our current efforts toward achieving optimal performance on GPU-based architectures, for pushing back the limitations that have prevented biologically realistic billion-atom objects to be fruitfully simulated, and for making large-scale simulations less expensive and easier to set up, run, and analyze. NAMD is distributed free of charge with its source code at www.ks.uiuc.edu.

Direct numerical simulations of turbulent flows using high-order asynchrony-tolerant schemes: Accuracy and performance

Direct numerical simulations (DNS) are an indispensable tool for understanding the fundamental physics of turbulent flows. Because of their steep increase in computational cost with Reynolds number (Rλ), well-resolved DNS are realizable only on massively parallel supercomputers, even at moderate Rλ. However, at extreme scales, the communications and synchronizations between processing elements (PEs) involved in current approaches become exceedingly expensive and are expected to be a major bottleneck to scalability. In order to overcome this challenge, we developed algorithms using the so-called Asynchrony-Tolerant (AT) schemes that relax communication and synchronization constraints at a mathematical level, to perform DNS of decaying and solenoidally forced compressible turbulence. Asynchrony is introduced using two approaches, one that avoids synchronizations and the other that avoids communications. These result in periodic and random delays, respectively, at PE boundaries. We show that both asynchronous algorithms accurately resolve the large-scale and small-scale motions of turbulence, including instantaneous and intermittent fields. We also show that in asynchronous simulations the communication time is a relatively smaller fraction of the total computation time, especially at large processor count, compared to standard synchronous simulations. As a consequence, we observe improved parallel scalability up to 262144 processors for both asynchronous algorithms.

Energy Efficient, Resource-Aware, Prediction Based VM Provisioning Approach for Cloud Environment

Over the past few decades, computing environments have progressed from a single-user milieu to highly parallel supercomputing environments, network of workstations (NoWs) and distributed systems, to more recently popular systems like grids and clouds. Due to its great advantage of providing large computational capacity at low costs, cloud infrastructures can be employed as a very effective tool, but due to its dynamic nature and heterogeneity, cloud resources consuming enormous amount of electrical power and energy consumption control becomes a major issue in cloud datacenters. This article proposes a comprehensive prediction-based virtual machine management approach that aims to reduce energy consumption by reducing active physical servers in cloud data centers. The proposed model focuses on three key aspects of resource management namely, prediction-based delay provisioning; prediction-based migration, and resource-aware live migration. The comprehensive model minimizes energy consumption without violating the service level agreement and provides the required quality of service. The experiments to validate the efficacy of the proposed model are carried out on a simulated environment, with varying server and user applications and parameter sizes.

In situ analysis and visualization of massively parallel simulations of transitional and turbulent flows

The increase of computational resources with the generalization of massively parallel supercomputers benefits to various fields of physics among which turbulence and fluid mechanics, making it possible to increase time and space accuracy and gain further knowledge in fundamental mechanisms. Parametric studies, high fidelity statistics, high resolutions, can be realized. However, this access poses many problems in terms of data management, analysis and visualization. Traditional workflow, consisting of writing raw data on disks and performing post-processing to extract physical quantities of interest, considerably slows down the analysis, if not becomes impossible, because of data transfer, storage and re-accessibility issues. This is particularly difficult when it comes to visualization. Usage has to be revisited to maintain consistency with the accuracy of the computation step and in this context, in situ processing is a promising approach. We developed an in situ analysis and visualization strategy with an hybrid method for transitional and turbulent flow analysis with a pseudo-spectral solver. It is shown to have a low impact on computational time with a reasonable increase of resource usage, while enriching data exploration. Large time sequences have been analyzed. This could not have been achieved with the traditional workflow. Moreover, computational steering has been performed with real-time adjustment of the simulations, thereby getting closer to a numerical experiment process.

The u-series: A separable decomposition for electrostatics computation with improved accuracy.

The evaluation of electrostatic energy for a set of point charges in a periodic lattice is a computationally expensive part of molecular dynamics simulations (and other applications) because of the long-range nature of the Coulomb interaction. A standard approach is to decompose the Coulomb potential into a near part, typically evaluated by direct summation up to a cutoff radius, and a far part, typically evaluated in Fourier space. In practice, all decomposition approaches involve approximations-such as cutting off the near-part direct sum-but it may be possible to find new decompositions with improved trade-offs between accuracy and performance. Here, we present the u-series, a new decomposition of the Coulomb potential that is more accurate than the standard (Ewald) decomposition for a given amount of computational effort and achieves the same accuracy as the Ewald decomposition with approximately half the computational effort. These improvements, which we demonstrate numerically using a lipid membrane system, arise because the u-series is smooth on the entire real axis and exact up to the cutoff radius. Additional performance improvements over the Ewald decomposition may be possible in certain situations because the far part of the u-series is a sum of Gaussians and can thus be evaluated using algorithms that require a separable convolution kernel; we describe one such algorithm that reduces communication latency at the expense of communication bandwidth and computation, a trade-off that may be advantageous on modern massively parallel supercomputers.