Search NASA⌕ Search

SEARCH · Search NASA

Results for “Parallelization efficiency”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,405 records · Page 78

Incentive Compatible Online Scheduling of Malleable Parallel Jobs with Individual Deadlines

We consider the online scheduling of malleable jobs on parallel systems, such as clusters, symmetric multiprocessing computers, and multi-core processor computers. Malleable jobs is a model of parallel processing in which jobs adapt to the number of processors assigned to them. This model permits the scheduler and resource manager to make more efficient use of the available resources. Each malleable job is characterized by arrival time, deadline, and value. If the job completes by its deadline, the user earns the payoff indicated by the value; otherwise, she earns a payoff of zero. The scheduling objective is to maximize the sum of the values of the jobs that complete by their associated deadlines. Complicating the matter is that users in the real world are rational and they will attempt to manipulate the scheduler by misreporting their jobs’ parameters if it benefits them to do so. To mitigate this behavior, we design an incentive compatible online scheduling mechanism. Incentive compatibility assures us that the users will obtain the maximum payoff only if they truthfully report their jobs’ parameters to the scheduler. Finally, we simulate and study the mechanism to show the effects of misreports on the cheaters and on the system.

99 GENERAL AND MISCELLANEOUS↗

GUNDAM: a toolkit for fast spatial correlation functions in galaxy surveys

ABSTRACT We describe the capabilities of a new software package to calculate two-point correlation functions (2PCFs) of large galaxy samples. The code can efficiently estimate 3D/projected/angular 2PCFs with a variety of statistical estimators and bootstrap errors, and is intended to provide a complete framework (including calculation, storage, manipulation, and plotting) to perform this type of spatial analysis with large redshift surveys. Gundam implements a very fast skip list/linked list algorithm that efficiently counts galaxy pairs and avoids the computation of unnecessary distances. It is several orders of magnitude faster than a naive pair counter, and matches or even surpass other advanced algorithms. The implementation is also embarrassingly parallel, making full use of multicore processors or large computational clusters when available. The software is designed to be flexible, user friendly and easily extensible, integrating optimized, well-tested packages already available in the astronomy community. Out of the box, it already provides advanced features such as custom weighting schemes, fibre collision corrections and 2D correlations. Gundam will ultimately provide an efficient toolkit to analyse the large-scale structure ‘buried’ in upcoming extremely large data sets generated by future surveys.

Donoso, E.↗

ION ACCELERATION IN NON-RELATIVISTIC ASTROPHYSICAL SHOCKS

We explore the physics of shock evolution and particle acceleration in non-relativistic collisionless shocks using hybrid simulations. We analyze a wide range of physical parameters relevant to the acceleration of cosmic rays (CRs) in astrophysical shock scenarios. We show that there are fundamental differences between high and low Mach number shocks in terms of the electromagnetic turbulence generated in the pre-shock zone; dominant modes are resonant with the streaming CRs in the low Mach number regime, while both resonant and non-resonant modes are present for high Mach numbers. Energetic power-law tails for ions in the downstream plasma account for up to 15% of the incoming upstream flow energy, distributed over {approx}5% of the particles in a power law with slope -2 {+-} 0.2 in energy. Quasi-parallel shocks with {theta} {<=} 45 Degree-Sign are good ion accelerators, while power laws are greatly suppressed for quasi-perpendicular shocks, {theta} > 45 Degree-Sign . The efficiency of conversion of flow energy into the energy of accelerated particles peaks at {theta} = 15 Degree-Sign -30 Degree-Sign and M{sub A} = 6, and decreases for higher Mach numbers, down to {approx}2% for M{sub A} = 31. Accelerated particles are produced by diffusive shock acceleration (DSA) and by shock drift acceleration (SDA) mechanisms, with the SDA contribution to the overall energy gain increasing with magnetic inclination. We also present a direct comparison between hybrid and fully kinetic particle-in-cell results at early times. In supernova remnant (SNR) shocks, particle acceleration will be significant for low Mach number quasi-parallel flows (M{sub A} < 30, {theta} < 45). This finding underscores the need for an effective magnetic amplification mechanism in SNR shocks.

79 ASTRONOMY AND ASTROPHYSICS↗

Pass-efficient methods for compression of high-dimensional turbulent flow data

The future of high-performance computing, specifically on future Exascale computers, will presumably see memory capacity and bandwidth fail to keep pace with data generated, for instance, from massively parallel partial differential equation (PDE) systems. Current strategies proposed to address this bottleneck entail the omission of large fractions of data, as well as the incorporation of in situ compression algorithms to avoid overuse of memory. To ensure that post-processing operations are successful, this must be done in a way that a sufficiently accurate representation of the solution is stored. Moreover, in situations where the input/output system becomes a bottleneck in analysis, visualization, etc., or the execution of the PDE solver is expensive, the number of passes made over the data must be minimized. In the interest of addressing this problem, this work focuses on the utility of pass-efficient, parallelizable, low-rank, matrix decomposition methods in compressing high-dimensional simulation data from turbulent flows. Additionally, a particular emphasis is placed on using coarse representation of the data – compatible with the PDE discretization grid – to accelerate the construction of the low-rank factorization. This includes the presentation of a novel single-pass matrix decomposition algorithm for computing the so-called interpolative decomposition. The methods are described extensively and numerical experiments on two turbulent channel flow data are performed. In the first (unladen) channel flow case, compression factors exceeding 400 are achieved while maintaining accuracy with respect to first- and second-order flow statistics. In the particle-laden case, compression factors of 100 are achieved and the compressed data is used to recover particle velocities. These results show that these compression methods can enable efficient computation of various quantities of interest in both the carrier and disperse phases.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Evaluation of High Rotor Pole Switched Reluctance Motors to Control Condenser Fans in a Commercial Refrigeration System

A high rotor pole switched reluctance motor (HRSRM) designed by Software Motor Corporation was evaluated for use as a condenser fan motor due to its reported high motor efficiency over a broad range of speeds and added advanced control potential. The project was split into two phases; the first of which was to determine the direct energy savings due to the increased motor efficiency, the second was to leverage the additional benefits of the energy efficiency measure (EEM) to create additional energy savings. This report only discusses results from the first phase of the project. Nine 1.5hp HRSRMs were installed parallel to the same number of legacy induction motors of a supermarket commercial refrigeration system located in Lakeside, Colorado. At this site, the 1.5 hp legacy motors had historically adjusted condenser fan speed from 200-870 rpm based on outdoor air temperature and compressor discharge pressure. This type of control is known as variable head pressure control (VHPC). These nine new motors were given a signal from the legacy variable frequency drive (VFD) to match the motor speed of the legacy motors. Determination of the energy savings of the EEM was performed over a two-month period, at which time a correlation between legacy and EEM power consumption was developed based on motor speed. Annual data that had been collected from 2016-2017 (before the two-month test period) was then leveraged to predict the annual energy savings of the EEM for a complete retrofit using both VHPC and constant fan speed control. Table 1 displays the quantitative results. The EEM saved 27% of total energy consumption of the condenser fan motors for constant fan speed control, while the energy savings increased to 32% when using VHPC. This was due to the improved motor efficiency of the HRSR motor at low speeds. While the constant speed fan control kept the motors running at 870 rpm, VHPC adjusted fan speed between 200-870rpm based on the outdoor air temperature and the compressor discharge pressure. These savings are assuming a direct replacement with no additional controls. Retrofitting a constant speed fan control system with HRSRMs and switching to VHPC would provide significant energy savings looking at the condenser alone. Additional considerations of the EEM were noted during the project as well. The HRSRM system made it possible to identify and remotely shutdown an individual motor when a piece of foam locked the fan and thereby the motor. With the traditional system, a locked rotor could have caused motor damage unless a breaker or fuse was activated. Furthermore, due to vibration complaints from the installation technicians, the ramp-up time of the HRSRM had to be increased to a minute. After implementing this change no complaints for excessive vibration were recorded.

15 GEOTHERMAL ENERGY↗

Batched matrix computations on hardware accelerators based on GPUs

Scientific applications require solvers that work on many small size problems that are independent from each other. At the same time, the high-end hardware evolves rapidly and becomes ever more throughput-oriented and thus there is an increasing need for an effective approach to develop energy-efficient, high-performance codes for these small matrix problems that we call batched factorizations. The many applications that need this functionality could especially benefit from the use of GPUs, which currently are four to five times more energy efficient than multicore CPUs on important scientific workloads. This study, consequently, describes the development of the most common, one-sided factorizations, Cholesky, LU, and QR, for a set of small dense matrices. The algorithms we present together with their implementations are, by design, inherently parallel. In particular, our approach is based on representing the process as a sequence of batched BLAS routines that are executed entirely on a GPU. Importantly, this is unlike the LAPACK and the hybrid MAGMA factorization algorithms that work under drastically different assumptions of hardware design and efficiency of execution of the various computational kernels involved in the implementation. Thus, our approach is more efficient than what works for a combination of multicore CPUs and GPUs for the problems sizes of interest of the application use cases. The paradigm where upon a single chip (a GPU or a CPU) factorizes a single problem at a time is not at all efficient in our applications’ context. We illustrate all of these claims through a detailed performance analysis. With the help of profiling and tracing tools, we guide our development of batched factorizations to achieve up to two-fold speedup and three-fold better energy efficiency as compared against our highly optimized batched CPU implementations based on MKL library. Finally, the tested system featured two sockets of Intel Sandy Bridge CPUs and we compared with a batched LU factorizations featured in the CUBLAS library for GPUs, we achieve as high as 2.5× speedup on the NVIDIA K40 GPU.

97 MATHEMATICS AND COMPUTING↗

Structure of a novel dodecaheme cytochrome c from Geobacter sulfurreducens reveals an extended 12 nm protein with interacting hemes.

Multiheme cytochromes c are important in electron transfer pathways in reduction of both soluble and insoluble Fe(III) by Geobacter sulfurreducens. We determined the crystal structure at 3.2 {angstrom} resolution of the first dodecaheme cytochrome c (GSU1996) along with its N-terminal and C-terminal hexaheme fragments at 2.6 and 2.15 {angstrom} resolution, respectively. The macroscopic reduction potentials of the full-length protein and its fragments were measured. The sequence of GSU1996 can be divided into four c{sub 7}-type domains (A, B, C and D) with homology to triheme cytochromes c{sub 7}. In cytochromes c{sub 7} all three hemes are bis-His coordinated, whereas in c{sub 7}-type domains the last heme is His-Met coordinated. The full-length GSU1996 has a 12 nm long crescent shaped structure with the 12 hemes arranged along a polypeptide to form a 'nanowire' of hemes; it has a modular structure. Surprisingly, while the C-terminal half of the protein consists of two separate c{sub 7}-type domains (C and D) connected by a small linker, the N-terminal half of the protein has two c{sub 7}-type domains (A and B) that form one structural unit. This is also observed in the AB fragment. There is an unexpected interaction between the hemes at the interface of domains A and B, which form a heme-pair with nearly parallel stacking of their porphyrin rings. The hemes adjacent to each other throughout the protein are within van der Waals distance which enables efficient electron exchange between them. For the first time, the structural details of c{sub 7}-type domains from one multiheme protein were compared.

59 BASIC BIOLOGICAL SCIENCES↗

Acceleration of Semiempirical QM/MM Methods through Message Passage Interface (MPI), Hybrid MPI/Open Multiprocessing, and Self-Consistent Field Accelerator Implementations

The strategy and implementation of scalable and efficient semiempirical (SE) QM/MM methods in CHARMM are described. The serial version of the code was first profiled to identify routines that required parallelization. Afterward, the code was parallelized and accelerated with three approaches. The first approach was the parallelization of the entire QM/MM routines, including the Fock matrix diagonalization routines, using the CHARMM message passage interface (MPI) machinery. In the second approach, two different self-consistent field (SCF) energy convergence accelerators were implemented using density and Fock matrices as targets for their extrapolations in the SCF procedure. In the third approach, the entire QM/MM and MM energy routines were accelerated by implementing the hybrid MPI/open multiprocessing (OpenMP) model in which both the task- and loop-level parallelization strategies were adopted to balance loads between different OpenMP threads. The present implementation was tested on two solvated enzyme systems (including <100 QM atoms) and an S N 2 symmetric reaction in water. The MPI version exceeded existing SE QM methods in CHARMM, which include the SCC-DFTB and SQUANTUM methods, by at least 4-fold. The use of SCF convergence accelerators further accelerated the code by ~12–35% depending on the size of the QM region and the number of CPU cores used. Although the MPI version displayed good scalability, the performance was diminished for large numbers of MPI processes due to the overhead associated with MPI communications between nodes. This issue was partially overcome by the hybrid MPI/OpenMP approach which displayed a better scalability for a larger number of CPU cores (up to 64 CPUs in the tested systems).

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Time-resolved RIXS experiment with pulse-by-pulse parallel readout data collection using X-ray free electron laser

Abstract Time-resolved resonant inelastic X-ray scattering (RIXS) is one of the developing techniques enabled by the advent of X-ray free electron laser (FEL). It is important to evaluate how the FEL jitter, which is inherent in the self-amplified spontaneous emission process, influences the RIXS measurement. Here, we use a microchannel plate (MCP) based Timepix soft X-ray detector to conduct a time-resolved RIXS measurement at the Ti L 3 -edge on a charge-density-wave material TiSe 2 . The fast parallel Timepix readout and single photon sensitivity enable pulse-by-pulse data acquisition and analysis. Due to the FEL jitter, low detection efficiency of spectrometer, and low quantum yield of RIXS process, we find that less than 2% of the X-ray FEL pulses produce signals, preventing acquiring sufficient data statistics while maintaining temporal and energy resolution in this measurement. These limitations can be mitigated by using future X-ray FELs with high repetition rates, approaching MHz such as the European XFEL in Germany and LCLS-II in the USA, as well as by utilizing advanced detectors, such as the prototype used in this study.

36 MATERIALS SCIENCE↗

Planck intermediate results: XXXV. Probing the role of the magnetic field in the formation of structure in molecular clouds

Within ten nearby (d < 450 pc) Gould belt molecular clouds we evaluate in this paper statistically the relative orientation between the magnetic field projected on the plane of sky, inferred from the polarized thermal emission of Galactic dust observed by Planck at 353 GHz, and the gas column density structures, quantified by the gradient of the column density, N H . The selected regions, covering several degrees in size, are analysed at an effective angular resolution of 10' FWHM, thus sampling physical scales from 0.4 to 40 pc in the nearest cloud. The column densities in the selected regions range from N H ≈ 10 21 to10 23 cm -2 , and hence they correspond to the bulk of the molecular clouds. The relative orientation is evaluated pixel by pixel and analysed in bins of column density using the novel statistical tool called “histogram of relative orientations”. Throughout this study, we assume that the polarized emission observed by Planck at 353 GHz is representative of the projected morphology of the magnetic field in each region, i.e., we assume a constant dust grain alignment efficiency, independent of the local environment. Within most clouds we find that the relative orientation changes progressively with increasing N H , from mostly parallel or having no preferred orientation to mostly perpendicular. In simulations of magnetohydrodynamic turbulence in molecular clouds this trend in relative orientation is a signature of Alfvénic or sub-Alfvénic turbulence, implying that the magnetic field is significant for the gas dynamics at the scales probed by Planck. Finally, we compare the deduced magnetic field strength with estimates we obtain from other methods and discuss the implications of the Planck observations for the general picture of molecular cloud formation and evolution.

79 ASTRONOMY AND ASTROPHYSICS↗

Advances in long pulse operation at high radio frequency power in Tore Supra

The lower hybrid current drive (LHCD) system of Tore Supra has been upgraded for long pulse operation at higher power (7–8 MW). The two launchers have coupled on plasma 3.8 MW and 2.7 MW separately. This new power capability allows extending the operational domain of Tore Supra for long pulses at higher current and density. 38 long (20 s –155 s) discharges with very low loop voltage (V{sub L} = 30-60 mV) were performed with combined LHCD (5-5.7 MW) and ICRH (1–3 MW) powers, with up to 1 GJ of injected energy. Higher LHCD efficiency, with respect to the previous long discharges, is reported. MHD stability of these discharges is very sensitive to the LHCD power and parallel wave index, in particular in the preforming phase. For theses evanescent loop voltage plasmas, the ICRH power, in excess of 1 MW, is found to have a beneficial effect on the MHD stability.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

PERPENDICULAR ION HEATING BY LOW-FREQUENCY ALFVEN-WAVE TURBULENCE IN THE SOLAR WIND

We consider ion heating by turbulent Alfven waves (AWs) and kinetic Alfven waves (KAWs) with wavelengths (measured perpendicular to the magnetic field) that are comparable to the ion gyroradius and frequencies {omega} smaller than the ion cyclotron frequency {Omega}. We focus on plasmas in which {beta} {approx}< 1, where {beta} is the ratio of plasma pressure to magnetic pressure. As in previous studies, we find that when the turbulence amplitude exceeds a certain threshold, an ion's orbit becomes chaotic. The ion then interacts stochastically with the time-varying electrostatic potential, and the ion's energy undergoes a random walk. Using phenomenological arguments, we derive an analytic expression for the rates at which different ion species are heated, which we test by simulating test particles interacting with a spectrum of randomly phased AWs and KAWs. We find that the stochastic heating rate depends sensitively on the quantity {epsilon} = {delta}v {sub {rho}/}v{sub perpendicular}, where v{sub perpendicular} (v {sub ||}) is the component of the ion velocity perpendicular (parallel) to the background magnetic field B {sub 0}, and {delta}v {sub {rho}} ({delta}B {sub {rho}}) is the rms amplitude of the velocity (magnetic-field) fluctuations at the gyroradius scale. In the case of thermal protons, when {epsilon} << {epsilon}{sub crit}, where {epsilon}{sub crit} is a constant, a proton's magnetic moment is nearly conserved and stochastic heating is extremely weak. However, when {epsilon}>{epsilon}{sub crit}, the proton heating rate exceeds half the cascade power that would be present in strong balanced KAW turbulence with the same value of {delta}v {sub {rho}}, and magnetic-moment conservation is violated even when {omega} << {Omega}. For the random-phase waves in our test-particle simulations, {epsilon}{sub crit} = 0.19. For protons in low-{beta} plasmas, {epsilon} {approx_equal} {beta}{sup -1/2{delta}}B{sub {rho}/}B {sub 0}, and {epsilon} can exceed {epsilon}{sub crit} even when {delta}B{sub {rho}/}B {sub 0} << {epsilon}{sub crit}. The heating is anisotropic, increasing v {sup 2}{sub perpendicular} much more than v {sup 2}{sub ||} when {beta} << 1. (In contrast, at {beta} {approx}> 1 Landau damping and transit-time damping of KAWs lead to strong parallel heating of protons.) At comparable temperatures, alpha particles and minor ions have larger values of {epsilon} than protons and are heated more efficiently as a result. We discuss the implications of our results for ion heating in coronal holes and the solar wind.

79 ASTRONOMY AND ASTROPHYSICS↗

Investigation of mode activity in NBI-heated experiments of Wendelstein 7-X

The 2018 operation phase (OP 1.2b) of the stellarator Wendelstein 7-X (W7-X) included, for the first time, neutral beam injection (NBI) to heat the plasma. Since the injection geometry at W7-X is not parallel, this generates both passing and trapped fast particles. During longer phases of NBI injection, with the primary purpose to study the heating efficiency of this system, Alfvén eigenmodes (AEs) were observed by a number of diagnostics, including the phase contrast imaging (PCI) system, the magnetic pick-up coils (Mirnov coils), and the soft x-ray multi-camera tomography system (XMCTS). Alfvén eigenmodes are of great interest for future fusion reactors as it has been shown that the resonant interaction of fast ions with self-excited AEs can lead to enhanced transport of fast ions and potentially to energy losses. This is especially true for so-called gap-modes, Alfvén eigenmodes with frequencies in gaps of the continuous spectrum, since they lack continuum damping. These modes are commonly known to be excited by fast ions, but other destabilizing mechanisms, e.g. the electron-pressure gradient are also possible. In this article we present a first analysis of the experimentally observed frequencies from the theoretical side. The calculation of shear Alfvén wave continua for selected cases and the assignment of observed frequencies to the gaps of the continuous spectra are presented. Using the ideal-MHD code CKA (Könies A. 2007 10th IAEA TM on Energetic Particles in Magnetic Confinement System), we find gap modes that match the experimental measurements in terms of the observed frequencies. We emphasize the crucial roles played by the coupling of sound and Alfvén waves as well as of the Doppler shift arising as a consequence of the radial electric field in W7-X. We employ the perturbative gyrokinetic code CKA-EUTERPE (Feh´er 2013 Simulation of the interaction between Alfv´en waves and fast particles), using a slowing-down distribution function for the fast ions as calculated by the Monte-Carlo particle following code ASCOT (Hirvijoki et al 2014 Comput. Phys. Commun. 185 1310–21) to assess the fast-ion drive. Here we find that the fast-ion drive is insufficient to overcome the background-plasma damping. The fact that unstable modes were observed experimentally may point to problems with the modelling or indicate the existence of other destabilizing mechanisms, e.g. associated with the electron-pressure gradient (Windisch et al 2017 Plasma Phys. Control. Fusion 59 105002) that sensitively depend on the profiles of the background plasma.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

MAS: A versatile Landau-fluid eigenvalue code for plasma stability analysis in general geometry

We have developed a new global eigenvalue code, Multiscale Analysis for plasma Stabilities (MAS), for studying plasma problems with wave toroidal mode number (n) and frequency (ω) in a broad range of interest in general tokamak geometry, based on a five-field Landau-fluid description of thermal plasmas. Beyond keeping the necessary plasma fluid response, we further retain the important kinetic effects including diamagnetic drift, ion finite Larmor radius, finite parallel electric field (E||), ion and electron Landau resonances in a self-consistent and non-perturbative manner without sacrificing the attractive efficiency in computation. The physical capabilities of the code are evaluated and examined in the aspects of both theory and simulation. In theory, the comprehensive Landau-fluid model implemented in MAS can be reduced to the well-known ideal MHD model, electrostatic ion-fluid model, and drift-kinetic model in various limits, which clearly delineates the physics validity regime. In simulation, MAS has been well benchmarked with theory and other gyrokinetic and kinetic-MHD hybrid codes in a manner of adopting the unified physical and numerical framework, which covers the kinetic Alfv\'en wave (KAW), ion sound wave (ISW), low-n kink, high-n ion temperature gradient mode (ITG) and kinetic ballooning mode (KBM). Moreover, MAS is successfully applied to model the Alfv\'en eigenmode (AE) activities in DIII-D discharge #159243, which faithfully captures the frequency sweeping of reversed shear Alfv\'en eigenmode (RSAE), the tunneling damping of toroidal Alfv\'en eigenmode (TAE), as well as the polarization characteristics of kinetic beta-induced Alfv\'en eigenmode (KBAE) and beta-induced Alfv\'en-acoustic eigenmode (BAAE) being consistent with former gyrokinetic theory and simulation. With respect to the key progress contributed to the community, MAS has the advantage of combining rich physics ingredients, realistic global geometry and high computation efficiency together for plasma stability analysis in linear regime.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Experimental Demonstration of a Controllable Electrostatic Molecular Beam Splitter

We experimentally demonstrate a controllable electrostatic beam splitter for guided ND{sub 3} molecules with a single Y-shaped charged wire and a homogeneous bias field generated by a charged metallic parallel-plate capacitor. We study the dependences of the splitting ratio R of the guided ND{sub 3} beam and its relative guiding efficiency {eta} on the voltage difference between two output arms of the splitter. The influences of the molecular velocity v and the cutting position L on the splitting ratio R are investigated as well, and the guiding and splitting dynamic processes of cold molecules are simulated. Our study shows that the splitting ratio R of our splitter can be conveniently adjusted from 10% to 90% by changing {Delta}U from -6 kV to +6 kV, and the simulated results are consistent with our experimental ones.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Performance Results on CPU/GPU Exascale Architectures for OMEGA: The Ocean Model for E3SM Global Applications

The US Department of Energy (DOE) conducts climate simulations on some of the world’s largest supercomputers. These exascale machines use heterogeneous architectures with both CPUs and GPUs, and scientific codes must adapt to make full use of this computing power. Los Alamos National Lab is developing Omega: The Ocean Model for E3SM Global Applications, which is specifically designed for modern exascale computers. It uses external libraries that have been optimized for a variety of architectures to run on different supercomputers. Omega is an unstructured-mesh ocean model based on TRiSK numerical methods. It will be the new ocean component of the DOE’s Energy Exascale Earth System Model (E3SM). The algorithms in Omega follow those of the current ocean component, MPAS-Ocean, but it will be written in C++ rather than Fortran to take advantage of the Kokkos performance portability library. Omega spatial operators are written as Kokkos kernels to run efficiently on both CPUs and GPUs. Work on Omega began in 2023 with a new C++ framework for unstructured mesh partitioning, halo exchanges, parallel IO, and Kokkos interfaces. The current version, Omega-0, is being developed to solve the shallow water equations and at present includes all of the tendency terms but not time stepping. Here we share the results of Omega-0 verification and performance testing. Verification includes unit tests implemented with CTest as well as convergence tests in Polaris, an in-house python package with a large suite of test problems. Performance tests compare simulations conducted on CPUs versus GPUs and across different architectures: tests are run on Frontier, which has AMD “Optimized 3rd Gen EPYC” CPUs and AMD MI250X GPUs, as well as Perlmutter, which is composed of AMD EPYC 7763 CPUs and NVIDIA A100 GPUs.

58 GEOSCIENCES↗

LHC Event Generation in the Exascale Era

MCFM is a dedicated Monte-Carlo simulation program for collider phenomenology at highest energies. Designed during the Tevatron era, it has successfully incorporated the latest developments needed for LHC precision calculations and remained on the forefront of collider phenomenology. The Fortran code includes interfaces to modern PDF and loop reduction libraries but has been unchanged structurally compared to the earlier versions. Parallel computing has been enabled using OpenMP and MPI. MCFM provides numerically highly stable one-loop amplitudes and superior phase-space efficiency, leading to excellent performance in NXLO calculations using jettiness or qT subtraction techniques for IR regularization.

Campbell, John [Fermilab]↗