Search NASA⌕ Search

SEARCH · Search NASA

Results for “high performance parallel computing”

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 991 records · Page 55
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

The Effect of Geometry, Spin, and Orbital Optimization in Achieving Accurate, Correlated Results for Iron–Sulfur Cubanes

Iron–sulfur clusters comprise an important functional motif in the catalytic centers of biological systems, capable of enabling important chemical transformations at ambient conditions. This remarkable capability derives from a notoriously complex electronic structure that is characterized by a high density of states that is sensitive to geometric changes. The spectral sensitivity to subtle geometric changes has received little attention from correlated, large active space calculations, owing partly to the exceptional computational complexity for treating these large and correlated systems accurately. Here, to provide insight into this aspect, we report the first Complete Active Space Self Consistent Field (CASSCF) calculations for different geometries of the [Fe(II/III) 4 S 4 (SMe) 4 ] -2 clusters using two complementary, correlated solvers: spin-pure Adaptive Sampling Configuration Interaction (ASCI) and Density Matrix Renormalization Group (DMRG). We find that the previously established picture of a double-exchange driven magnetic structure, with minute energy gaps (<1 mHa) between consecutive spin states, has a weak dependence on the underlying geometry. However, the spin gap between the singlet and the spin state 2S + 1 = 19, corresponding to a maximal number of Fe-d electrons being unpaired and of parallel spin, is strongly geometry dependent, changing by a factor of 3 upon slight deformations that are still within biologically relevant parameters. The CASSCF orbital optimization procedure, using active spaces as large as 86 electrons in 52 orbitals, was found to reduce this gap compared to typical mean-field orbital approaches. Our results show the need for performing large active space calculations to unveil the challenging electronic structure of these complex catalytic centers and should serve as accurate starting points for fully correlated treatments upon inclusion of dynamical correlation outside the active space.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2021

Characterizing Machine Learning I/O Workloads on Leadership Scale HPC Systems

High performance computing (HPC) is no longer solely limited to traditional workloads such as simulation and modeling. With the increase in the popularity of machine learning (ML) and deep learning (DL) technologies, we are observing that an increasing number of HPC users are incorporating ML methods into their workflow and scientific discovery processes, across a wide spectrum of science domains such as biology, earth science, and physics. This gives rise to a diverse set of I/O patterns than the traditional checkpoint/restart-based HPC I/O behavior. The details of the I/O characteristics of such ML I/O workloads have not been studied extensively for large-scale leadership HPC systems. This paper aims to fill that gap by providing an in-depth analysis to gain an understanding of the I/O behavior of ML I/O workloads using darshan - an I/O characterization tool designed for lightweight tracing and profiling. We study the darshan logs of more than 23, 000 HPC ML I/O jobs over a time period of one year running on Summit - the second-fastest supercomputer in the world. This paper provides a systematic I/O characterization of ML I/O jobs running on a leadership scale supercomputer to understand how the I/O behavior differs across science domains and the scale of workloads, and analyze the usage of parallel file system and burst buffer by ML I/O workloads.

Paul, Arnab↗
DOE OSTI2023

Code modernization strategies for short-range non-bonded molecular dynamics simulations

Modern HPC systems are increasingly relying on greater core counts and wider vector registers. Thus, applications need to be adapted to fully utilize these hardware capabilities. One class of applications that can benefit from this increase in parallelism are molecular dynamics simulations. In this paper, we describe our efforts at modernizing the ESPResSo++ simulation package for molecular dynamics by restructuring its particle data layout for efficient memory accesses and applying vectorization techniques to benefit the calculation of short-range non-bonded forces, which results in an overall three times speedup and serves as a baseline for further optimizations. We also implement fine-grained parallelism for multi-core CPUs through HPX, a C++ runtime system which uses lightweight threads and an asynchronous many-task approach to maximize concurrency. Our goal is to evaluate the performance of an HPX-based approach compared to the bulk-synchronous MPI-based implementation. This requires the introduction of an additional layer to the domain decomposition scheme that defines the task granularity. On spatially inhomogeneous systems, which impose a corresponding load-imbalance in traditional MPI-based approaches, we demonstrate that by choosing an optimal task size, the efficient work-stealing mechanisms of HPX can overcome the overhead of communication resulting in an overall 1.4 times speedup compared to the baseline MPI version.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Productive Programming of Distributed Systems with the SHAD C++ Library

High-performance computing (HPC) is often perceived as a matter of making large-scale systems (e.g., clusters) run as fast as possible, regardless the required programming effort. However, the idea of "bringing HPC to the masses" has recently emerged. Inspired by this vision, we have designed SHAD, the Scalable High-performance Algorithms and Data-structures library. SHAD is open source software, written in C++, for C++ developers. Unlike other HPC libraries for distributed systems, which rely on SPMD models, SHAD adopts a shared-memory programming abstraction, to make C++ programmers feel at home. Underneath, SHAD manages tasking and data-movements, moving the computation where data resides and taking advantage of asynchrony to tolerate network latency. At the bottom of his stack, SHAD can interface with multiple runtime systems: this not only improves developer’s productivity, by hiding the complexity of such software and of the underlying hardware, but also greatly enhance code portability. Thanks to its abstraction layers, SHAD can indeed target different systems, ranging from laptops to HPC clusters, without any need for modifying the user-level code. We have prototyped and open-sourced the implementation of (a subset of) the C++ standard library (STL) targeting multi-node HPC clusters. Our work allows plain STL-based C++ code to scale on HPC systems, with no need for rewriting the code to exploit the complex hardware. SHAD is available under Apache v2 License at https://github.com/pnnl/SHAD. In this paper we overview the design of the SHAD library, depicting its main components: runtime systems abstractions for tasking; parallel and distributed data-structures; STL-compliant interfaces and algorithms.

Castellana, Vito G.↗
DOE OSTI2025

Fragme∩t: An Open‐Source Framework for Multiscale Quantum Chemistry Based on Fragmentation

Fragment-based quantum chemistry offers a means to circumvent the nonlinear computational scaling of conventional electronic structure calculations, by partitioning a large calculation into smaller subsystems then considering the many-body interactions between them. Variants of this approach have been used to parameterize classical force fields and machine learning potentials, applications that benefit from interoperability between quantum chemistry codes. However, there is a dearth of software that provides interoperability yet is purpose-built to handle the combinatorial complexity of fragment-based calculations. To fill this void we introduce “Fragme∩t”, an open-source software application that provides a tool for community validation of fragment-based methods, a platform for developing new approximations, and a framework for analyzing many-body interactions. Fragme∩t includes algorithms for automatic fragment generation and structure modification, and for distance- and energy-based screening of the requisite subsystems. Checkpointing, database management, and parallelization are handled internally and results are archived in a portable database. Interfaces to various quantum chemistry engines are easy to write and exist already for Q-Chem, PySCF, xTB, Orca, CP2K, MRCC, Psi4, NWChem, GAMESS, and MOPAC. Applications reported here demonstrate parallel efficiencies around 96% on more than 1000 processors but also showcase that the code can handle large-scale protein fragmentation using only workstation hardware, all with a codebase that is designed to be usable by non-experts. Fragme∩t conforms to modern software engineering best practices and is built upon well established technologies including Python, SQLite, and Ray. The source code is available under the Apache 2.0 license.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2021

FROSch Preconditioners for Land Ice Simulations of Greenland and Antarctica

Numerical simulations of Greenland and Antarctic ice sheets involve the solution of large-scale highly nonlinear systems of equations on complex shallow geometries. This work is concerned with the construction of Schwarz preconditioners for the solution of the associated tangent problems, which are challenging for solvers mainly because of the strong anisotropy of the meshes and wildly changing boundary conditions that can lead to poorly constrained problems on large portions of the domain. Here, two-level GDSW (Generalized Dryja–Smith–Widlund) type Schwarz preconditioners are applied to different land ice problems, i.e., a velocity problem, a temperature problem, as well as the coupling of the former two problems. We employ the MPI-parallel implementation of multi-level Schwarz preconditioners provided by the package FROSch (Fast and Robust Schwarz)from the Trilinos library. The strength of the proposed preconditioner is that it yields out-of-the-box scalable and robust preconditioners for the single physics problems. To our knowledge, this is the first time two-level Schwarz preconditioners are applied to the ice sheet problem and a scalable preconditioner has been used for the coupled problem. The pre-conditioner for the coupled problem differs from previous monolithic GDSW preconditioners in the sense that decoupled extension operators are used to compute the values in the interior of the sub-domains. Several approaches for improving the performance, such as reuse strategies and shared memory OpenMP parallelization, are explored as well. In our numerical study we target both uniform meshes of varying resolution for the Antarctic ice sheet as well as non uniform meshes for the Greenland ice sheet are considered. We present several weak and strong scaling studies confirming the robustness of the approach and the parallel scalability of the FROSch implementation. Among the highlights of the numerical results are a weak scaling study for up to 32 K processor cores (8 K MPI-ranks and 4 OpenMP threads) and 566 M degrees of freedom for the velocity problem as well as a strong scaling study for up to 4 K processor cores (and MPI-ranks) and 68 M degrees of freedom for the coupled problem.

58 GEOSCIENCES↗
DOE OSTI2022

High-order multirate explicit time-stepping schemes for the baroclinic-barotropic split dynamics in primitive equations

In order to treat the multiple time scales of ocean dynamics in an efficient manner, the baroclinic-barotropic splitting technique has been widely used for solving the primitive equations for ocean modeling. Based on the framework of strong stability-preserving Runge-Kutta approach, we propose two high-order multirate explicit time-stepping schemes (SSPRK2-SE and SSPRK3-SE) for the resulting split system in this paper. The proposed schemes allow for a large time step to be used for the three-dimensional baroclinic (slow) mode and a small time step for the two-dimensional barotropic (fast) mode, in which each of the two mode solves just need to satisfy their respective CFL conditions for numerical stability. Specifically, at each time step, the baroclinic velocity is first computed by advancing the baroclinic mode and fluid thickness of the system with the large time-step and the assistance of some intermediate approximations of the barotropicmode obtained by substepping with the small time step; then the barotropic velocity is corrected by using the small time step to re-advance the barotropic mode un-der an improved barotropic forcing produced by interpolation of the forcing terms from the preceding baroclinic mode solves; lastly, the fluid thickness is updated by coupling the baroclinic and barotropic velocities. Additionally, numerical inconsistencies on the discretized sea surface height caused by the mode splitting are relieved via a reconciliation process with carefully calculated flux deficits. Here, two benchmark tests from the “MPAS-Ocean” platform are carried out to numerically demonstrate the performance and parallel scalability of the proposed SSPRK-SE schemes.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

Delamination-informed lifecycle decisions: A dielectric and machine learning framework for composite sorting and recycling

Composite materials are widely used in aerospace, marine, and automotive sectors due to their high strength-to-weight ratio and durability. However, their long-term reliability can be compromised by damage accumulation. Specifically, delamination initiation serves as a precursor to structural failure, which is often difficult to detect during damage inspection. Identifying and sorting delamination initiation in samples not only increases operational safety while providing critical information for end-of-life decisions, which influences both the service life extension value and the efficiency of fiber extraction during recycling. This research addresses two challenges: (1) developing a nondestructive, ex-situ framework to sort composite materials based on damage severity, particularly delamination, and (2) understanding how damage in composites influences resin removal during pyrolysis. Both experimental work and finite element analysis were performed to predict critical stress levels that are associated with delamination onset. Based on these results, three loading levels 50 %, 75 %, and 90 % of maximum stress, were selected for controlled experiments, generating composite samples with varying extents of damage for machine learning model training. Microscopic imaging of these samples confirmed the damage progression from matrix cracking to delamination, validating the computational predictions. We explored supervised machine learning using dielectric measurements to classify damage states. Preliminary results show an artificial neural network can identify early delamination which is a potential precursor to failure, with 94.44 % accuracy on our dataset. A parallel investigation into the effect of damage severity on pyrolysis recycling showed that heavily delaminated samples required significantly less energy for comparable matrix removal than undamaged samples.

dielectric variables↗
DOE OSTI2022

FROSch Preconditioners for Land Ice Simulations of Greenland and Antarctica

Numerical simulations of Greenland and Antarctic ice sheets involve the solution of large-scale highly nonlinear systems of equations on complex shallow geometries. This work is concerned with the construction of Schwarz preconditioners for the solution of the associated tangent problems, which are challenging for solvers mainly because of the strong anisotropy of the meshes and wildly changing boundary conditions that can lead to poorly constrained problems on large portions of the domain. In this study, two-level generalized Dryja--Smith--Widlund (GDSW)--type Schwarz preconditioners are applied to different land ice problems, i.e., a velocity problem, a temperature problem, as well as the coupling of the former two problems. We employ the message passing interface (MPI)--parallel implementation of multilevel Schwarz preconditioners provided by the package FROSch (fast and robust Schwarz) from the Trilinos library. The strength of the proposed preconditioner is that it yields out-of-the-box scalable and robust preconditioners for the single physics problems. To the best of our knowledge, this is the first time two-level Schwarz preconditioners have been applied to the ice sheet problem and a scalable preconditioner has been used for the coupled problem. The preconditioner for the coupled problem differs from previous monolithic GDSW preconditioners in the sense that decoupled extension operators are used to compute the values in the interior of the subdomains. Additionally, several approaches for improving the performance, such as reuse strategies and shared memory OpenMP parallelization, are explored. In our numerical study we target both uniform meshes of varying resolution for the Antarctic ice sheet as well as nonuniform meshes for the Greenland ice sheet. We present several weak and strong scaling studies confirming the robustness of the approach and the parallel scalability of the FROSch implementation. Among the highlights of the numerical results are a weak scaling study for up to 32,K processor cores (8,K MPI ranks and 4 OpenMP threads) and 566,M degrees of freedom for the velocity problem as well as a strong scaling study for up to 4,K processor cores (and MPI ranks) and 68,M degrees of freedom for the coupled problem.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2017

Surrogate Models for TREAT Transient Calculations

The Transient Reactor Test Facility (TREAT) at Idaho National Laboratory (INL) is currently preparing for resumption of operation after more than 20 years in a long-term standby state. In parallel with preparation for operations and a new transient testing program, the DOENE Advanced Modeling and Simulation program (NEAMS) has been supporting the development of three-dimensional transient multi-physics modeling capabilities within the reactor physics application MAMMOTH. Recently, the MAMMOTH reactor physics team has begun development of uncertainty and sensitivity analysis for the TREAT core. This research will be used to support the analysis of reactor operation and experiment planning. Traditional approaches, i.e., Monte Carlo simulations, for such tasks require a large number of model evaluations under both steady state and transient conditions. However, exhaustive sets of sampling runs are simply not practical for TREAT since evaluations of each model can take a substantial amount of time. A solution to this problem is to build a surrogate model employing a smaller number of model evaluations. The surrogate model is designed to provide an accurate and fast approximation to the detailed TREAT model. In this manner, the surrogate model can be used in place of the TREAT transient model for uncertainty and sensitivity analysis at a fraction of the computational cost. The high dimensional model representation (HDMR) method is employed here for approximating the responses of interest in terms of functions of lower dimensions. This method has been implemented in RAVEN and all TREAT transient calculations are performed using the multiphysics code MAMMOTH. In addition, a code interface has also been developed in order to couple RAVEN with MAMMOTH. The surrogate model is then tested and compared for TREAT startup testing program Transient Test no. 15 core, and its accuracy is measured by its ability to estimate the mean, coefficient of variance, skewness, excess kurtosis and probability density function of the outputs of Transient Test no. 15 model, as well as root mean square error metric.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
DOE OSTI2022

Pseudonymization at Scale: OLCF’s Summit Usage Data Case Study

The analysis of vast amounts of data and the processing of complex computational jobs have traditionally relied upon high performance computing (HPC) systems, which offer reliable and efficient management of large-scale computational and data resources. Understanding these analyses’ needs is paramount for designing solutions that can lead to better science, and similarly, understanding the characteristics of the user behavior on those systems is important for improving user experiences on HPC systems. A common approach to gathering data about user behavior is to extract workload characteristics from system log data available only to system administrators. Recently at Oak Ridge Leadership Computing Facility (OLCF), however, we unveiled user behavior about the Summit supercomputer by collecting data from a user’s point of view with ordinary Unix commands.In this paper, we discuss the process, challenges, and lessons learned while preparing this dataset for publication and submission to an open data challenge. The original dataset contains personal identifiable information (PII) about the users of OLCF which needed be masked prior to publication, and we determined that anonymization, which scrubs PII completely, destroyed too much of the structure of the data to be interesting for the data challenge. We instead chose to pseudonymize the dataset, which reduced the linkability of the dataset to the users’ identities. Pseudonymization is significantly more computationally expensive than anonymization, and the size of our dataset, which is approximately 175 million lines of raw text, necessitated the development of a parallelized workflow that could be reused on different HPC machines. We demonstrate the scaling behavior of the workflow on two leadership class HPC systems at OLCF, and we show that we were able to bring the overall makespan time from an impractical 20+ hours on a single node down to around 2 hours. As a result of this work, we release the entire pseudonymized dataset and make the workflows and source code publicly available.

Maheshwari, Ketan↗
DOE OSTI2018

Parallel Ellipsoidal Perfectly Matched Layers for Acoustic Helmholtz Problems on Exterior Domains

Exterior acoustic problems occur in a wide range of applications, making the finite element analysis of such problems a common practice in the engineering community. Various methods for truncating infinite exterior domains have been developed, including absorbing boundary conditions, infinite elements, and more recently, perfectly matched layers (PML). PML are gaining popularity due to their generality, ease of implementation, and effectiveness as an absorbing boundary condition. PML formulations have been developed in Cartesian, cylindrical, and spherical geometries, but not ellipsoidal. In addition, the parallel solution of PML formulations with iterative solvers for the solution of the Helmholtz equation, and how this compares with more traditional strategies such as infinite elements, has not been adequately investigated. In this study, we present a parallel, ellipsoidal PML formulation for acoustic Helmholtz problems. To faciliate the meshing process, the ellipsoidal PML layer is generated with an on-the-fly mesh extrusion. Though the complex stretching is defined along ellipsoidal contours, we modify the Jacobian to include an additional mapping back to Cartesian coordinates in the weak formulation of the finite element equations. This allows the equations to be solved in Cartesian coordinates, which is more compatible with existing finite element software, but without the necessity of dealing with corners in the PML formulation. Herein we also compare the conditioning and performance of the PML Helmholtz problem with infinite element approach that is based on high order basis functions. On a set of representative exterior acoustic examples, we show that high order infinite element basis functions lead to an increasing number of Helmholtz solver iterations, whereas for PML the number of iterations remains constant for the same level of accuracy. Finally, this provides an additional advantage of PML over the infinite element approach.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2026

Spherical tokamak physics research in preparation for the operation of NSTX-U

The National Spherical Torus Experiment Upgrade (NSTX-U) is preparing to resume operation, representing a crucial step toward realizing compact, cost-effective fusion pilot plants. In advance of this, extensive modeling and data analysis have been conducted to advance the physics basis for low-aspect-ratio, high-performance plasma regimes, focusing on three core objectives: confinement and stability, power and particle handling, and steady-state operation. Significant progress has been made in understanding the electron temperature flattening in high-β plasmas, which is shown to be driven by a complex interplay of magnetohydrodynamic instabilities (e.g. non-resonant infernal modes), fast-ion-driven Alfvén eigenmodes, and electron and ion-scale micro-instabilities, particularly Kinetic Ballooning Modes (KBMs), whose destabilization is strongly dependent on parallel magnetic field fluctuations (δB ∥ ). Furthermore, a new gyrokinetic critical pedestal model was developed, accurately predicting pedestal structure by identifying KBMs as the primary stability limit, offering a critical constraint for future high-confinement scenarios. To address the challenge of high heat flux, novel liquid lithium plasma-facing components were modeled. The analysis confirmed that lithium vapor shielding is a self-regulating mechanism for heat mitigation, while also emphasizing that strong main ion parallel flow is essential to minimize core lithium contamination. Finally, progress toward steady-state operation was anchored by developing the required physics basis and control tools. This includes predictive modeling for reversed magnetic shear sustainment, demonstrating that magnetic island-induced bootstrap current reduction is negligible in STs, and advancing real-time control and disruption avoidance capabilities. The development of high-speed surrogate models (e.g. MMMNet) provides computationally efficient tools vital for non-inductive scenario optimization and integrated, low-disruptivity operations planned for NSTX-U.

NSTX-U↗
DOE OSTI2017

Surrogate Models for TREAT Transient Calculations

The Transient Reactor Test Facility (TREAT) at Idaho National Laboratory (INL) is currently preparing for resumption of operation after more than 20 years in a long-term standby state. In parallel with preparation for operations and a new transient testing program, the DOE-NE Advanced Modeling and Simulation program (NEAMS) has been supporting the development of three-dimensional transient multi-physics modeling capabilities within the reactor physics application MAMMOTH [1]. Recently, the MAMMOTH reactor physics team has begun development of uncertainty and sensitivity analysis for the TREAT core. This research will be used to support the analysis of reactor operation and experiment planning. Traditional approaches, i.e., Monte Carlo simulations, for such tasks require a large number of model evaluations under both steady state and transient conditions. However, exhaustive sets of sampling runs are simply not practical here since evaluations of each model can take a substantial amount of time. A solution to this problem is to build a surrogate model employing a smaller number of model evaluations. The surrogate model is designed to provide an accurate and fast approximation to the detailed TREAT model. In this manner, the surrogate model can be used in place of the TREAT transient model for uncertainty and sensitivity analysis at a fraction of the computational cost. The high dimensional model representation (HDMR) [2] method is employed here for approximating the responses of interest in terms of functions of lower dimensions. This method has been implemented in RAVEN [3] and all TREAT transient calculations are performed using the multiphysics code MAMMOTH. In addition, a code interface has also been developed in order to couple RAVEN with MAMMOTH. The surrogate model is then tested and compared for the MAMMOTH model of the TREAT startup testing program Transient Test #15 core [4], and its accuracy is measured by its ability to estimate the mean, coefficient of variance, skewness, excess kurtosis and probability density function of the outputs of Transient Test #15 model, as well as root mean square error metric.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2018

High-Performance Molecular Dynamics Simulation for Biological and Materials Sciences: Challenges of Performance Portability, Dallas, TX, November 2018.

Highly-optimized parallel molecular dynamics programs have allowed researchers to achieve ground-breaking results in biological and materials sciences. This type of performance has come at the expense of portability: a significant effort is required for performance optimization on each new architecture. Using a metric that emphasizes speedup, we assess key accelerating programming components of four different best-performing molecular dynamics programs - GROMACS, NAMD, LAMMPS and CP2K-- each having a particular scope of application, for contribution to performance and for portability. We use builds with and without these components, tested on HPC systems. We also analyze the code-bases to determine compliance with portability recommendations. We find that for all four programs, the contributions of the non-portable components to speed are essential to the programs' performances; without them we see a reduction in time-to-solution of a magnitude that is insufferable to domain scientists. This characterizes the performance efficiency that must be approached for good performance portability on a programmatic level, suggesting solutions to this difficult problem, which should come from developers, industry and funding institutions, and possibly new research in programming languages.

Computer Science↗
DOE OSTI2017

Task-based Parallel Computation of the Density Matrix in Quantum-based Molecular Dynamics using Graph Partitioning

Quantum-based molecular dynamics (QMD) is a highly accurate and transferable method for material science simulations. However, the time scales and system sizes accessible to QMD are typically limited to picoseconds and a few hundred atoms. These constraints arise due to expensive self-consistent ground-state electronic structure calculations that can often scale cubically with the number of atoms. Linearly scaling methods depend on computing the density matrix P from the Hamiltonian matrix H by exploiting the sparsity in both matrices. The second-order spectral projection (SP2) algorithm is an O(N) algorithm that computes P with a sequence of 40 - 50 matrix-matrix multiplications. In this paper, we present task-based implementations of a recently developed data-parallel graph-based approach to the SP2 algorithm, G-SP2. We represent the density matrix P as an undirected graph and use graph partitioning techniques to divide the computation into smaller independent tasks. The partitions thus obtained are generally not of equal size and give rise to undesirable load imbalances in standard MPI-based implementations. This load-balancing challenge can be mitigated by dynamically scheduling parallel computations at runtime using task-based programming models. We develop task-based implementations of the data-parallel G-SP2 algorithm using both Intel's Concurrent Collections (CnC) as well as the Charm++ programming model and evaluate these implementations for future use. Scaling and performance results of our implementations are investigated for representative segments of QMD simulations for solvated protein systems containing more than 10,000 atoms.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

A Performance Portable, Fully Implicit Landau Collision Operator with Batched Linear Solvers

Modern accelerators use hierarchical parallel programming models that enable massive multithreading within a processing element (PE), with multiple PEs per device driven by traditional processes. Batching is a technique for exposing PE-level parallelism in algorithms that have traditionally run on MPI processes or multiple threads within a single process. Opportunities for batching arise in, for example, kinetic discretizations of magnetized plasmas where collisions are advanced in velocity space at each spatial point independently. This paper builds on previous work on a high-performance, fully nonlinear, Landau collision operator by batching the linear solver, as well as batching the spatial point problems and adding new support for multiple grids for multiscale, multispecies problems. An anisotropic relaxation verification test that agrees well with previously published results and analytical models is presented. The performance results from NVIDIA A100 and AMD MI250X nodes are presented with hardware utilization analysis for each architecture. Finally, the entire implicit Landau operator time advance is implemented in Kokkos for performance portability, running entirely on the device and is available in the PETSc numerical library.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Communication-Avoiding and Memory-Constrained Sparse Matrix-Matrix Multiplication at Extreme Scale

Sparse matrix-matrix multiplication (SpGEMM) is a widely used kernel in various graph, scientific computing and machine learning algorithms. In this paper, we consider SpGEMMs performed on hundreds of thousands of processors generating trillions of nonzeros in the output matrix. Distributed SpGEMM at this extreme scale faces two key challenges: (1) high communication cost and (2) inadequate memory to generate the output. Furthermore, we address these challenges with an integrated communication-avoiding and memory-constrained SpGEMM algorithm that scales to 262,144 cores (more than 1 million hardware threads) and can multiply sparse matrices of any size as long as inputs and a fraction of output fit in the aggregated memory. As we go from 16,384 cores to 262,144 cores on a Cray XC40 supercomputer, the new SpGEMM algorithm runs 10x faster when multiplying large-scale protein-similarity matrices.

97 MATHEMATICS AND COMPUTING↗