Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Parallel Supercomputers”

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 559 records · Page 31

A parallel strategy for density functional theory computations on accelerated nodes

Using the Löwdin orthonormalization of tall-skinny matrices as a proxy-app for wavefunction-based Density Functional Theory solvers, we investigate a distributed memory parallel strategy focusing on Graphics Processing Unit (GPU)-accelerated nodes as available on some of the top ranked supercomputers at the present time. Here we present numerical results in the strong limit regime, as it is particularly relevant for First-Principles Molecular Dynamics. We also examine how matrix product-based iterative solvers provide a competitive alternative to dense eigensolvers on GPUs, allowing to push the strong scaling limit of these computations to a larger number of distributed tasks. Our strategy, which relies on replicated Gram matrices and efficient collective communications using the NCCL library, leads to a time-to-solution under 0.5 s for the Löwdin orthonormalization of a tall-skinny matrix of 3000 columns on Summit at Oak Ridge Leadership Facility (OLCF). Given the similarity in computational operations between one iteration of a DFT solver and this proxy-app, this shows the possibility of solving accurately the DFT equations well under a minute for 3000 electronic wave functions, and thus perform First-Principles molecular dynamics of physical systems much larger than traditionally solved on CPU systems.

97 MATHEMATICS AND COMPUTING↗

HPDR: High-Performance Portable Scientific Data Reduction Framework

The rapid growth in scientific data generation is outpacing advancements in computing systems necessary for efficient storage, transfer, and analysis, particularly in the context of exascale computing. With the deployment of first-generation exascale computing systems and next-generation experimental facilities, this gap is widening and necessitates effective data reduction techniques to manage enormous data volumes. Over the past decade, various data reduction methods, including lossless compression, error-controlled lossy compression, and data refactoring, have been developed to accelerate I/O in scientific workflows. Despite significant reductions in data volume, these methods introduce considerable computational overhead, which can become the new bottleneck in data processing. To mitigate this, GPU-accelerated data reduction algorithms have been introduced. However, challenges remain in their integration into exascale workflows, including limited portability across different GPU architectures, substantial memory transfer overhead, and reduced scalability on dense multi-GPU systems. To address these challenges, we propose HPDR, a high-performance and portable data reduction framework. HPDR is designed to enable the execution of state-of-the-art reduction algorithms across diverse processor architectures while reducing memory transfer overhead to 2.3 % of the original, resulting in up to 3.5× faster throughput compared to existing solutions. It also achieves up to 96% of the theoretical speedup in multi-GPU settings. In addition, evaluations on accelerating I/O operations at scale up to 1,024 nodes of the Frontier supercomputer demonstrate that HPDR can achieve up to 103 TB/s reduction throughput, providing up to 4× acceleration in parallel I/O performance compared to existing data reduction routines. This work highlights the potential of HPDR to significantly enhance data reduction efficiency in exascale computing environments.

Chen, Jieyang [University of Oregon]↗

Designing Scalable Distributed Memory Models: A Case Study

Supercomputers complexity continue to rapidly increase, requiring novel software solutions to aid in addressing the extreme abundance of parallelism, restrictive power constraints, and fault tolerance requirements. While industrial standards like MPI and OpenMP have undergone extensive efforts to prepare for future systems [1, 2], ne grain asynchronous runtimes have demonstrated an innate agility in extracting performance which is otherwise arduous to attain with current execution models

Landwehr, Joshua B.↗

Numerical Verification of the Power Transfer and Wakefield Coupling in the Clic Two-Beam Accelerator

The Compact Linear Collider (CLIC) provides a path to a multi-TeV accelerator to explore the energy frontier of High Energy Physics. Its two-beam accelerator (TBA) concept envisions complex 3D structures, which must be modeled to high accuracy so that simulation results can be directly used to prepare CAD drawings for machining. The required simulations include not only the fundamental mode properties of the accelerating structures but also the Power Extraction and Transfer Structure (PETS), as well as the coupling between the two systems. Time-domain simulations will be performed to understand pulse formation, wakefield damping, fundamental power transfer and wakefield coupling in these structures. Applying SLAC's parallel finite element code suite, these large-scale problems will be solved on some of the largest supercomputers available. The results will help to identify potential issues and provide new insights on the design, leading to further improvements on the novel two-beam accelerator scheme.

43 PARTICLE ACCELERATORS↗

Simulation-based Performance Analysis and Tuning for a Two-level Directly Connected System

Hardware and software co-design is becoming increasingly important due to complexities in supercomputing architectures. Simulating applications before there is access to the real hardware can assist machine architects in making better design decisions that can optimize application performance. At the same time, the application and runtime can be optimized and tuned beforehand. BigSim is a simulation-based performance prediction framework designed for these purposes. It can be used to perform packet-level network simulations of parallel applications using existing parallel machines. In this paper, we demonstrate the utility of BigSim in analyzing and optimizing parallel application performance for future systems based on the PERCS network. We present simulation studies using benchmarks and real applications expected to run on future supercomputers. Future petascale systems will have more than 100,000 cores, and we present simulations at that scale.

97 MATHEMATICS AND COMPUTING↗

Particle simulation on heterogeneous distributed supercomputers

We describe the implementation and performance of a three dimensional particle simulation distributed between a Thinking Machines CM-2 and a Cray Y-MP. These are connected by a combination of two high-speed networks: a high-performance parallel interface (HIPPI) and an optical network (UltraNet). This is the first application to use this configuration at NASA Ames Research Center. We describe our experience implementing and using the application and report the results of several timing measurements. We show that the distribution of applications across disparate supercomputing platforms is feasible and has reasonable performance. In addition, several practical aspects of the computing environment are discussed.

Becker, Jeffrey C.↗

Massively parallel electrical conductivity imaging of the subsurface: Applications to hydrocarbon exploration

Three-dimensional (3D) geophysical imaging is now receiving considerable attention for electrical conductivity mapping of potential offshore oil and gas reservoirs. The imaging technology employs controlled source electromagnetic (CSEM) and magnetotelluric (MT) fields and treats geological media exhibiting transverse anisotropy. Moreover when combined with established seismic methods, direct imaging of reservoir fluids is possible. Because of the size of the 3D conductivity imaging problem, strategies are required exploiting computational parallelism and optimal meshing. The algorithm thus developed has been shown to scale to tens of thousands of processors. In one imaging experiment, 32,768 tasks/processors on the IBM Watson Research Blue Gene/L supercomputer were successfully utilized. Over a 24 hour period we were able to image a large scale field data set that previously required over four months of processing time on distributed clusters based on Intel or AMD processors utilizing 1024 tasks on an InfiniBand fabric. Electrical conductivity imaging using massively parallel computational resources produces results that cannot be obtained otherwise and are consistent with timeframes required for practical exploration problems.

54 ENVIRONMENTAL SCIENCES↗

Unstructured grids on SIMD torus machines

Unstructured grids lead to unstructured communication on distributed memory parallel computers, a problem that has been considered difficult. Here, we consider adaptive, offline communication routing for a SIMD processor grid. Our approach is empirical. We use large data sets drawn from supercomputing applications instead of an analytic model of communication load. The chief contribution of this paper is an experimental demonstration of the effectiveness of certain routing heuristics. Our routing algorithm is adaptive, nonminimal, and is generally designed to exploit locality. We have a parallel implementation of the router, and we report on its performance.

Bjorstad, Petter E.↗

Parallel interior-point solver for block-structured nonlinear programs on SIMD/GPU architectures

Here, we investigate how to port the standard interior-point method to new exascale architectures for block-structured nonlinear programs with state equations. Computationally, we decompose the interior-point algorithm into two successive operations: the evaluation of the derivatives and the solution of the associated Karush-Kuhn-Tucker (KKT) linear system. Our method accelerates both operations using two levels of parallelism. First, we distribute the computations on multiple processes using coarse parallelism. Second, each process uses SIMD/GPU accelerators locally to accelerate the operations using fine-grained parallelism. The KKT system is reduced by eliminating the inequalities and the state variables from the corresponding equations. We demonstrate our method's capability on the supercomputer Polaris, a testbed for the future exascale Aurora system. Each node is equipped with four GPUs, a setup amenable to our two-level approach. Our experiments on the stochastic optimal power flow problem show that the reduction method is 50x faster than the sparse linear solver HSL MA57 running in serial on the CPU, and 6x faster than Pardiso running in parallel on CPU on the same number of processes.

97 MATHEMATICS AND COMPUTING↗

Simulation of Unsteady Combustion in a Ramjet Engine Using a Highly Parallel Computer

Combustion instability in ramjets is a complex phenomenon that involve nonlinear interaction between acoustic waves, vortex motion and unsteady heat release in the combustor. To numerically simulate this 3-D, transient phenomenon, enormous computer resources (time, memory and disk storage) are required. Although current generation vector supercomputers are capable of providing adequate resources for simulations of this nature, their high cost and limited availability, makes such machines less than satisfactory for routine use. The primary focus of this study is to assess the feasibility of using highly parallel computer systems as a cost-effective alternative for conducting such unsteady flow simulations. Towards this end, a large-eddy simulation model for combustion instability was implemented on the Intel iPSC/860 and a careful study was conducted to determine the benefits and the problems associated with the use of such machines for transient simulations. Details of this study along with the results obtained from the unsteady combustion simulations carried out on the iPSC/860 are discussed in this paper.

Menon, Suresh↗

Task parallel sensitivity analysis and parameter estimation of groundwater simulations through the SALSSA framework

The Support Architecture for Large-Scale Subsurface Analysis (SALSSA) provides an extensible framework, sophisticated graphical user interface, and underlying data management system that simplifies the process of running subsurface models, tracking provenance information, and analyzing the model results. Initially, SALSSA supported two styles of job control: user directed execution and monitoring of individual jobs, and load balancing of jobs across multiple machines taking advantage of many available workstations. Recent efforts in subsurface modelling have been directed at advancing simulators to take advantage of leadership class supercomputers. We describe two approaches, current progress, and plans toward enabling efficient application of the subsurface simulator codes via the SALSSA framework: automating sensitivity analysis problems through task parallelism, and task parallel parameter estimation using the PEST framework.

99 GENERAL AND MISCELLANEOUS↗

Distributed memory, GPU accelerated Fock construction for hybrid, Gaussian basis density functional theory

With the growing reliance of modern supercomputers on accelerator-based architecture such a graphics processing units (GPUs), the development and optimization of electronic structure methods to exploit these massively parallel resources has become a recent priority. While significant strides have been made in the development GPU accelerated, distributed memory algorithms for many modern electronic structure methods, the primary focus of GPU development for Gaussian basis atomic orbital methods has been for shared memory systems with only a handful of examples pursing massive parallelism. Here in this work, we present a set of distributed memory algorithms for the evaluation of the Coulomb and exact exchange matrices for hybrid Kohn–Sham DFT with Gaussian basis sets via direct density-fitted (DF-J-Engine) and seminumerical (sn-K) methods, respectively. The absolute performance and strong scalability of the developed methods are demonstrated on systems ranging from a few hundred to over one thousand atoms using up to 128 NVIDIA A100 GPUs on the Perlmutter supercomputer.

97 MATHEMATICS AND COMPUTING↗

GPU-accelerated Tersoff potentials for massively parallel Molecular Dynamics simulations

The Tersoff potential is one of the empirical many-body potentials that has been widely used in simulation studies at atomic scales. Unlike pair-wise potentials, the Tersoff potential involves three-body terms, which require much more arithmetic operations and data dependency. In this contribution, we have implemented the GPU-accelerated version of several variants of the Tersoff potential for LAMMPS, an open-source massively parallel Molecular Dynamics code. Compared to the existing MPI implementation in LAMMPS, the GPU implementation exhibits a better scalability and offers a speedup of 2.2X when run on 1000 compute nodes on the Titan supercomputer. On a single node, the speedup ranges from 2.0 to 8.0 times, depending on the number of atoms per GPU and hardware configurations. The most notable features of our GPU-accelerated version include its design for MPI/accelerator heterogeneous parallelism, its compatibility with other functionalities in LAMMPS, its ability to give deterministic results and to support both NVIDIA CUDA- and OpenCL-enabled accelerators. Our implementation is now part of the GPU package in LAMMPS and accessible for public use.

Computer Science↗

Communication Characterization and Optimization of Applications Using Topology-Aware Task Mapping on Large Supercomputers

On large supercomputers, the job scheduling systems may assign a non-contiguous node allocation for user applications depending on available resources. With parallel applications using MPI (Message Passing Interface), the default process ordering does not take into account the actual physical node layout available to the application. This contributes to non-locality in terms of physical network topology and impacts communication performance of the application. In order to mitigate such performance penalties, this work describes techniques to identify suitable task mapping that takes the layout of the allocated nodes as well as the application's communication behavior into account. During the first phase of this research, we instrumented and collected performance data to characterize communication behavior of critical US DOE (United States - Department of Energy) applications using an augmented version of the mpiP tool. Subsequently, we developed several reordering methods (spectral bisection, neighbor join tree etc.) to combine node layout and application communication data for optimized task placement. We developed a tool called mpiAproxy to facilitate detailed evaluation of the various reordering algorithms without requiring full application executions. This work presents a comprehensive performance evaluation (14,000 experiments) of the various task mapping techniques in lowering communication costs on Titan, the leadership class supercomputer at Oak Ridge National Laboratory.

97 MATHEMATICS AND COMPUTING↗

Communication Characterization and Optimization of Applications Using Topology-Aware Task Mapping on Large Supercomputers

On large supercomputers, the job scheduling systems may assign a non-contiguous node allocation for user applications depending on available resources. With parallel applications using MPI (Message Passing Interface), the default process ordering does not take into account the actual physical node layout available to the application. This contributes to non-locality in terms of physical network topology and impacts communication performance of the application. In order to mitigate such performance penalties, this work describes techniques to identify suitable task mapping that takes the layout of the allocated nodes as well as the application's communication behavior into account. During the first phase of this research, we instrumented and collected performance data to characterize communication behavior of critical US DOE (United States - Department of Energy) applications using an augmented version of the mpiP tool. Subsequently, we developed several reordering methods (spectral bisection, neighbor join tree etc.) to combine node layout and application communication data for optimized task placement. We developed a tool called mpiAproxy to facilitate detailed evaluation of the various reordering algorithms without requiring full application executions. This work presents a comprehensive performance evaluation (14,000 experiments) of the various task mapping techniques in lowering communication costs on Titan, the leadership class supercomputer at Oak Ridge National Laboratory.

Sreepathi, Sarat↗

Parallel scaling analysis for explicit solid dynamics in ALEGRA

Weak scaling studies were performed for the explicit solid dynamics component of the ALEGRA code on two Cray supercomputer platforms during the period 2012-2015, involving a production-oriented hypervelocity impact problem. Results from these studies are presented, with analysis of the performance, scaling, and throughput of the code on these machines. The analysis demonstrates logarithmic scaling of the average CPU time per cycle up to core counts on the order of 10,000. At higher core counts, variable performance is observed, with significant upward excursions in compute time from the logarithmic trend. However, for core counts less than 10,000, the results show a 3 × improvement in simulation throughput, and a 2 × improvement in logarithmic scaling. This improvement is linked to improved memory performance on the Cray platforms, and to significant improvements made over this period to the data layout used by ALEGRA.

97 MATHEMATICS AND COMPUTING↗

Many-Core Acceleration of a Discrete Ordinates Transport Mini-App at Extreme Scale, In: ISC High Performance 2016: High Performance Computing

Time-dependent deterministic discrete ordinates transport codes are an important class of application which provide significant challenges for large, many-core systems. One such challenge is the large memory capacity needed by the solve step, which requires us to have a scalable solution in order to have enough node-level memory to store all the data. In our previous work, we demonstrated the first implementation which showed a significant performance benefit for single node solves using GPUs. In this paper we extend our work to large problems and demonstrate the scalability of our solution on two Petascale GPU-based supercomputers: Titan at Oak Ridge and Piz Daint at CSCS. Our results show that our improved node-level parallelism scheme scales just as well across large systems as previous approaches when using the tried and tested KBA domain decomposition technique. We validate our results against an improved performance model which predicts the runtime of the main ‘sweep’ routine when running on different hardware, including CPUs or GPUs.

Computer Science↗