Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “high performance space 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 415 records · Page 23

Using Parameter Sweep in WaterTAP to Analyze New Water Treatment Technologies

We describe a powerful and generalized parameter sweep tool in this report that was originally developed to analyze the performance of existing and novel water treatment models being developed in WaterTAP. Since WaterTAP is built upon IDAES and Pyomo, the parameter sweep tool can be used to systematically explore and debug the behavior of most Pyomo and IDAES numerical models. In order to enable meaningful analyses, the parameter sweep tool has been designed with the following features: 1) Model flexibility: The parameter sweep tool does not enforce any restrictions on the types of models that can be used with it. As long as a Pyomo model can be solved and the parameter is active and mutable, the tool only needs functions that describe how to run the model, the sweep parameters, and the output quantities of interest. 2) Flexible sampling: The parameter sweep tool has inbuilt functions to generate samples from a random distribution or a multidimensional Euclidean space. Furthermore, the users have to ability to supply samples generated from a tool of their choice. 3) Multiple sweep types: A user can choose from one of 3 types of parameter sweeps depending on their needs. 4) Detailed outputs: Outputs generated by the parameter sweep tool can be stored in detailed H5 file or user-friendly CSV files for post processing. 5) Parallel computing: The parameter sweep supports shared and distributed memory parallel computing to enable the use of high performance computers (HPC) for large-scale analyses. 6) Modular: The parameter sweep tool is self-contained and can easily be integrated within an outer-loop analysis or as desired by the user. 7) Ease of use: The tool is well documented and a simple sweep can be easily executed by following the online documentation in a few lines of code. We demonstrate the use of the parameter sweep tool on a simple water treatment system from the WaterTAP repository and show its parallel scaling performance on an Apple laptop and NREL's Eagle HPC. The parameter sweep tool is actively being used with models currently being developed within WaterTAP and we expect its use to grow beyond it to other IDAES and Pyomo models.

97 MATHEMATICS AND COMPUTING↗

A Tailored Convolutional Neural Network for Nonlinear Manifold Learning of Computational Physics Data Using Unstructured Spatial Discretizations

In this work, we propose a nonlinear manifold learning technique based on deep convolutional autoencoders that is appropriate for model order reduction of physical systems in complex geometries. Convolutional neural networks have proven to be highly advantageous for compressing data arising from systems demonstrating a slow-decaying Kolmogorov n-width. However, these networks are restricted to data on structured meshes. Unstructured meshes are often required for performing analyses of real systems with complex geometry. Our custom graph convolution operators based on the available differential operators for a given spatial discretization effectively extend the application space of deep convolutional autoencoders to systems with arbitrarily complex geometry that are typically discretized using unstructured meshes. We propose sets of convolution operators based on the spatial derivative operators for the underlying spatial discretization, making the method particularly well suited to data arising from the solution of partial differential equations. We demonstrate the method using examples from heat transfer and fluid mechanics and show better than an order of magnitude improvement in accuracy over linear methods.

97 MATHEMATICS AND COMPUTING↗

Evaluating the potential of disaggregated memory systems for HPC applications

Summary Disaggregated memory is a promising approach that addresses the limitations of traditional memory architectures by enabling memory to be decoupled from compute nodes and shared across a data center. Cloud platforms have deployed such systems to improve overall system memory utilization, but performance can vary across workloads. High‐performance computing (HPC) is crucial in scientific and engineering applications, where HPC machines also face the issue of underutilized memory. As a result, improving system memory utilization while understanding workload performance is essential for HPC operators. Therefore, learning the potential of a disaggregated memory system before deployment is a critical step. This paper proposes a methodology for exploring the design space of a disaggregated memory system. It incorporates key metrics that affect performance on disaggregated memory systems: memory capacity, local and remote memory access ratio, injection bandwidth, and bisection bandwidth, providing an intuitive approach to guide machine configurations based on technology trends and workload characteristics. We apply our methodology to analyze thirteen diverse workloads, including AI training, data analysis, genomics, protein, fusion, atomic nuclei, and traditional HPC bookends. Our methodology demonstrates the ability to comprehend the potential and pitfalls of a disaggregated memory system and provides motivation for machine configurations. Our results show that eleven of our thirteen applications can leverage injection bandwidth disaggregated memory without affecting performance, while one pays a rack bisection bandwidth penalty and two pay the system‐wide bisection bandwidth penalty. In addition, we also show that intra‐rack memory disaggregation would meet the application's memory requirement and provide enough remote memory bandwidth.

Ding, Nan↗

A theoretical and experimental benchmark study of core-excited states in nitrogen

The high resolution near edge X-ray absorption fine structure spectrum of nitrogen displays the vibrational structure of the core-excited states. This makes nitrogen well suited for assessing the accuracy of different electronic structure methods for core excitations. We report high resolution experimental measurements performed at the SOLEIL synchrotron facility. These are compared with theoretical spectra calculated using coupled cluster theory and algebraic diagrammatic construction theory. The coupled cluster singles and doubles with perturbative triples model known as CC3 is shown to accurately reproduce the experimental excitation energies as well as the spacing of the vibrational transitions. In conclusion, the computational results are also shown to be systematically improved within the coupled cluster hierarchy, with the coupled cluster singles, doubles, triples, and quadruples method faithfully reproducing the experimental vibrational structure.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Scaling comb-driven resonator-based DWDM silicon photonic links to multi-Tb/s in the multi-FSR regime

The use of chip-based micro-resonator Kerr frequency combs in conjunction with dense wavelength-division multiplexing (DWDM) enables massively parallel intensity-modulated direct-detection data transmission with low energy consumption. Resonator-based modulators and filters used in such systems can limit the number of usable wavelength channels due to practical constraints on the maximum achievable free spectral range (FSR). In this work, we introduce the design of multi-Tb/s comb-driven resonator-based silicon photonic links by leveraging the multi-FSR regime. We demonstrate the viability of the link architecture with yield estimates that are supported by extensive wafer-scale measurements of 704 micro-resonators fabricated in a commercial complementary metal–oxide–semiconductor foundry. We show that a 2.80 Tb/s link is realizable with a ≥6 σ yield (∼99.999%), and that aggregate bandwidths of 3.76 Tb/s and 4.72 Tb/s are possible if yield targets are relaxed (3 σ and 1 σ , respectively). All designs represent a 1.94−3.28× boost to aggregate link bandwidth while maintaining BER≤10 −10 performance, with a theoretical bandwidth of 10.51 Tb/s being possible for sufficiently robust resonators. We use high-speed BER measurements to inform co-optimization of data rate and aggressor spacing ( λ ag ), limiting any additional loss-based power penalties to off-resonance insertion loss (IL) and routing loss. This work demonstrates that, through the multi-FSR regime, there is a clear path toward Kerr comb-driven ultra-broadband, high bandwidth silicon photonic links that can support next-generation data centers and high-performance computers.

James, Aneek (ORCID:0000000262527807)↗

Permanent Magnets Featuring Heavy Main Group Elements for Magnetic Anisotropy

Permanent magnets are the functional component of electric motors and generators found in numerous renewable energy applications. To improve energy conversion in such applications, we require fundamentally new magnets that generate higher magnetic flux per volume while retaining the properties conferred by rare-earth elements incorporated into current technologies. We hypothesize that by engendering a covalent interaction between two elements, we can access a new regime of magnetic materials where the two components of a magnetic moment—spin and orbital angular momentum—come from two separate atoms to form a complete magnetic moment. Our previous research utilized high-pressure conditions to discover two new candidate materials ideal for assessing this hypothesis. The first material, FeBi 2 , enables the study of an unprecedented solid state metal-metal bonding interaction. The second, MnBi 2 represents the second member of the promising Mn–Bi family known for its magnetic properties. Importantly, MnBi 2 is isostructural to FeBi 2 . Together these chemically simple but magnetically rich materials provide an elegant platform for elucidating fundamental design principals of magnetic anisotropy while inspiring the synthesis of new magnetic materials. More generally, solid-state chemistry remains a synthetic black box. To fully harness the potential of such new materials, it is vital to create a window into both the structure and the properties of our high-pressure materials. One missing area of importance to energy science is high-pressure magnetometry. To this end, we assessed the magnetic structure of these materials through two objectives: (1) we performed magnetometry measurements at high-pressures, including performing high-pressure X-ray Magnetic Circular Dichroism (XMCD) experiments on these materials that are only synthesized at high-pressure and cannot be recovered to ambient conditions and (2) we targeted the recovery of these materials using dynamic compression approaches, starting with laser shock techniques. Future directions include expanding our focus to ternary phase space, employing high-throughput computational approaches to direct our search for magnetic materials across the Periodic Table. This award was originated at Northwestern, and reopened as a new grant at MIT. This is the close out for Northwestern.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Streaming Analytics for Anomaly Detection in Large-Scale Data

Anomalous behavior poses serious risks to assured performance and reliability of complex, high-consequence systems. For spaceborne assets and their state-of-health (SOH) telemetry, the challenges of high-dimensional data of varying data types are compounded by computational limitations from size, weight, and power (SWaP) constraints as well as data availability. Automated anomaly detection methods tend to perform poorly under these constraints, while current operational approaches can introduce delays in response time due to the manual, retrospective processes for understanding system failures. As a result, presently deployed space systems, and those deployed in the near future, face situations where mission operations might be delayed or only be able to operate under degraded capabilities. Here, we examine a near-term lightweight solution that provides real-time detection capabilities for rare events and assess state-of-the-art anomaly detection techniques against real SOH telemetry from space platforms. This report describes our methodology and research, which could support more automated capabilities for comprehensive space operations as well as for other resource-constrained edge applications.

97 MATHEMATICS AND COMPUTING↗

High-Performance Multi-Mode Ptychography Reconstruction on Distributed GPUs

Abstract—Ptychography is an emerging imaging technique that is able to provide wavelength-limited spatial resolution from spec-imen with extended lateral dimensions. As a scanning microscopy method, a typical two-dimensional image requires a number of data frames. As a diffraction-based imaging technique, the real-space image has to be recovered through iterative reconstruction algorithms. Due to these two inherent aspects, a ptychographic reconstruction is generally a computation-intensive and time-consuming process, which limits the throughput of this method. We report an accelerated version of the multi-mode difference map algorithm for ptychography reconstruction using multiple distributed GPUs. This approach leverages available scientific computing packages in Python, including mpi4py and PyCUDA, with the core computation functions implemented in CUDA C. We find that interestingly even with MPI collective communications, the weak scaling in the number of GPU nodes can still remain nearly constant. Most importantly, for realistic diffraction measurements, we observe a speedup ranging from a factor of 10 to 103 depending on the data size, which reduces the reconstruction time remarkably from hours to typically about 1 minute and is thus critical for real-time data processing and visualization.

97 MATHEMATICS AND COMPUTING↗

Accelerating high-order continuum kinetic plasma simulations using multiple GPUs

Kinetic plasma simulations solve the Vlasov-Poisson or Vlasov-Maxwell equations to evolve scalar-variable distribution functions in position-velocity phase space and vector-variable electromagnetic fields in configuration space. The immense computational cost of evolving high-dimensional variables, and their large number of degrees of freedom, often limits the utility of continuum kinetic simulations and presents a challenge when it comes to accurately simulating real-world physical phenomena. To address this challenge, we present techniques that accelerate and minimize the computational work required for a scalable Vlasov-Poisson solver. We show theoretical hardware compute and communication bounds for solving a fourth-order finite-volume Vlasov-Poisson system. These bounds are then used to inform and evaluate the design of performance portable algorithms for a multiple graphics processing unit (GPU) accelerated version of the Vlasov-Poisson solver VCK-CPU [1]. We demonstrate that the multi-GPU Vlasov solver implementation, VCK-GPU, simultaneously minimizes required inter-process data transfer while also being bounded by the machine network performance limits. This results in an overall strong scaling speedup per timestep of up to 40x in three-dimensional phase space (one position, two velocity coordinates) and 54x in four dimensional phase space (two position, two velocity coordinates) and a 341x increase in simulation throughput of the GPU accelerated code over the existing CPU code. The GPU code is also able to weak scale up to 256 compute nodes and 1024 GPUs. In conclusion, we demonstrate that the improved compute performance enables exploring configurations which were previously computationally infeasible, including resolving fine-scale distribution function filamentation and multi-species dynamics with realistic electron-proton mass ratios.

Continuum kinetics↗

UCX: An Open Source Framework for HPC Network APIs and Beyond, In: 2015 IEEE 23rd Annual Symposium on High-Performance Interconnects

This paper presents Unified Communication X (UCX), a set of network APIs and their implementations for high throughput computing. UCX comes from the combined effort of national laboratories, industry, and academia to design and implement a high-performing and highly-scalable network stack for next generation applications and systems. UCX design provides the ability to tailor its APIs and network functionality to suit a wide variety of application domains and hardware. We envision these APIs to satisfy the networking needs of many programming models such as Message Passing Interface (MPI), OpenSHMEM, Partitioned Global Address Space (PGAS) languages, task-based paradigms and I/O bound applications. To evaluate the design we implement the APIs and protocols, and measure the performance of overhead-critical network primitives fundamental for implementing many parallel programming models and system libraries. Our results show that the latency, bandwidth, and message rate achieved by the portable UCX prototype is very close to that of the underlying driver. With UCX, we achieved a message exchange latency of 0.89 us, a bandwidth of 6138.5 MB/s, and a message rate of 14 million messages per second. As far as we know, this is the highest bandwidth and message rate achieved by any network stack (publicly known) on this hardware.

Computer Science↗

Exploring the Use of Novel Spatial Accelerators in Scientific Applications

Driven by the need to find alternative accelerators which can viably replace GPUs in next-generation Supercomputing systems, this paper proposes a methodology to enable agile application/hardware co-design. The application-first methodology provides the ability to come up with design of accelerators while working with real-world workloads, available accelerators, and system software. The iterative design process targets a set of kernels in a workload for performance estimates that can prune the design space for later phases of detailed architectural evaluations. To this effect, in this paper, a novel data-parallel device model is introduced that simulates the latency of performance-sensitive operations in an accelerator including data transfers and kernel computation using multi-core CPUs. The use of off-the-shelf simulators, such as pre-RTL simulator Aladdin or multiple tools available for exploring the design of deep neural network accelerators (e.g., Timeloop) is demonstrated for evaluation of various accelerator designs using applications with realistic inputs. Examples of multiple device configurations that are instantiable in a system are explored to evaluate the performance benefit of deploying novel accelerators. The proposed device is integrated with a programming model and system software to potentially explore the impacts of high-level programming languages/compilers and low-level effects such as task scheduling on multiple accelerators. We analyze our methodology for a set of applications that represent high-performance computing (HPC) and graph analytics. The applications include a computational chemistry kernel realized using tensor contractions, triangle counting, GraphSAGE and Breadth-first Search. These applications include kernels such as dense matrix-dense matrix multiplication, sparse matrix-spare matrix multiplication, and sparse matrix-dense vector multiplication. Our results indicate potential performance benefits and insights for system design by including accelerators that realize these kernels along-side general purpose accelerators.

AI, codesign, Accelerated Computing, Modeling and ↗

A novel low-rank matrix completion approach to estimate missing entries in Euclidean distance matrix

A Euclidean distance matrix (EDM) is a table of distance-square between points on a k-dimensional Euclidean space, with applications in many fields (e.g., engineering, geodesy, economics, genetics, biochemistry, and psychology). A problem that often arises is the absence (or uncertainty) of some EDM elements. In many situations, only a subset of all pairwise distances is available and it is desired to have some procedure to estimate the missing distances. In this paper, we address the problem of missing data in EDM through low-rank matrix completion techniques. We exploit the fact that the rank of a EDM is at most k+2 and does not depend on the number of points, which is, in general, much bigger then k. We use a singular value decomposition approach that considers the rank of the matrix to be completed and computes, in each iteration, a parameter that controls the convergence of the method. After performing a number of computational experiments, we could observe that our proposal was able to recover, with high precision, random EDMs with more than 1000 points and up to 98% of missing data in few minutes. In addition, our method required a smaller number of iterations when compared to other competitive state-of-art technique.

97 MATHEMATICS AND COMPUTING↗

Real-Time Spatio-Temporal Twice Whitening for MIMO Energy Detector

While many techniques exist for local spectrum sensing of a primary user, each represents a computationally demanding task to secondary user receivers. In software-defined radio, computational complexity lengthens the time for a cognitive radio to recognize changes in the transmission environment. This complexity is even more significant for spatially multiplexed receivers, e.g., in SIMO and MIMO, where the spatio-temporal data sets grow in size with the number of antennae. Limits on power and space for the processor hardware further constrain SDR performance. In this report, we discuss improvements in spatio-temporal twice whitening (STTW) for real-time local spectrum sensing by demonstrating a form of STTW well suited for MIMO environments. We implement STTW on the Coherent Logix hx3100 processor, a multicore processor intended for low-power, high-throughput software-defined signal processing. These results demonstrate how coupling the novel capabilities of emerging multicore processors with algorithmic advances can enable real-time, software-defined processing of large spatio-temporal data sets.

97 MATHEMATICS AND COMPUTING↗

Computationally Predicted High-Throughput Free-Energy Phase Diagrams for the Discovery of Solid-State Hydrogen Storage Reactions

The design of multinary solid-state material systems that undergo reversible phase changes via changes in temperature and pressure provides a potential means of safely storing hydrogen. However, fully mapping the stabilities of known or newly targeted compounds relative to competing phases at reaction conditions has previously required many stringent experiments or computationally demanding calculations of each compound’s change in Gibbs energy with respect to temperature, G(T). Here, we have extended the approach of constructing chemical potential phase diagrams based on ΔG f (T) to enable the analysis of phase stability at non-zero temperatures. We first performed density functional theory calculations to compute the formation enthalpies of binary, ternary, and quaternary compounds within several compositional spaces of current interest for solid-state hydrogen storage. Temperature effects on solid compound stability were then accounted for using our recently introduced machine learned descriptor for the temperature-dependent contribution G δ (T) to the Gibbs energy G(T). From these Gibbs energies, we evaluated each compound’s stability relative to competing compounds over a wide range of conditions and show using chemical potential and composition phase diagrams that the predicted stable phases and H2 release reactions are consistent with experimental observations. This demonstrates that our approach rapidly computes the thermochemistry of hydrogen release reactions for compounds at sufficiently high accuracy relative to experiment to provide a powerful framework for analyzing hydrogen storage materials. This framework based on G(T) enables the accelerated discovery of active materials for a variety of technologies that rely on solid-state reactions involving these materials.

08 HYDROGEN↗

Physics vs structure: A systematic benchmark of learning strategies for multi-zone building thermal dynamics

Recent advances in physics-informed and data-driven machine learning promise improved thermal models for advanced building control, yet there is limited quantitative evidence on when added physics structure and architectural complexity are beneficial. Here, this work presents a systematic benchmark of five representative system identification methods for modeling multi-zone building thermal dynamics: linear state-space models, multi-layer perceptrons, neural state-space models, neural ordinary differential equations, and physically-consistent neural networks. The methods are evaluated across multiple data regimes and zone coupling strategies. Using a high-fidelity multi-zone commercial building emulator, we examine short-term and long-term prediction accuracy, computational efficiency, and ease of development. Our results reveal critical trade-offs between prediction performance, model complexity, and physical consistency. We demonstrate that decoupled, nonlinear black-box models consistently outperform coupled physics-constrained architectures in both predictive accuracy and out-of-distribution robustness in majority of the test cases for the building type considered in the study. Our findings quantify the cost of complexity in building thermal modeling and provide concrete, actionable, scenario-based guidelines for selecting model classes for control-oriented applications.

Building thermal modeling↗

Accelerating scientific discoveries through data-driven innovations

Developing artificial intelligence (AI) and machine learning (ML) methods that can accelerate scientific discoveries and advance science has become one of the important research directions for the AI/ML research community. It has been gaining increasing attention from researchers in diverse scientific areas, including biomedical science, materials science, climate science, physics, chemistry, and many others. Data-driven AI/ML innovations to enable reliable predictions and optimal decision making for scientific discoveries face several critical challenges, among which are high system complexity, large search space, incomplete knowledge, and small data, all of which demand novel strategies to effectively address them. Meeting these challenges and thereby accelerating scientific discoveries and industrial innovations, calls for research that can take full advantage of the latest advances in AI/ML to integrate data-driven techniques with scientific knowledge and is able to execute them in modern high-performance computing (HPC) environments at scale. This Patterns special collection "Accelerating scientific discoveries through data-driven innovations" features articles that showcase the promising roles of AI/ML and data-driven modeling in accelerating scientific discoveries and may inspire the next wave of data-driven innovations in various scientific domains.

97 MATHEMATICS AND COMPUTING↗

Simulating X-ray Absorption Spectra with Linear-Response Density Cumulant Theory

We present a new approach for simulating X-ray absorption spectra based on linear-response density cumulant theory (LR-DCT) [Copan, A. V.; Sokolov, A. Yu. J. Chem. Theory Comput. 2018, 14, 4097–4108]. Our new method combines the LR-ODC-12 formulation of LR-DCT with core- valence separation approximation (CVS) that allows us to efficiently access high-energy core-excited states. We describe our computer implementation of the CVS-approximated LRODC-12 method (CVS-ODC-12) and benchmark its performance by comparing simulated X-ray absorption spectra to those obtained from experiment for several small molecules. Our results demonstrate that the CVS-ODC-12 method shows good agreement with experiment for relative spacings between transitions and their intensities, but the excitation energies are systematically overestimated. When compared to results from excited-state coupled cluster methods with single and double excitations, the CVS-ODC-12 method shows a similar performance for intensities and peak separations, while coupled cluster spectra are less shifted, relative to experiment. An important advantage of CVS-ODC-12 is that its excitation energies are computed by diagonalizing a Hermitian matrix, which enables efficient computation of transition intensities.

Peng, Ruojing↗

Simulation of adiabatic quantum computing for molecular ground states

Quantum computation promises to provide substantial speedups in many practical applications with a particularly exciting one being the simulation of quantum many-body systems. Adiabatic state preparation (ASP) is one way that quantum computers could recreate and simulate the ground state of a physical system. In this paper, we explore a novel approach for classically simulating the time dynamics of ASP with high accuracy and with only modest computational resources via an adaptive sampling configuration interaction scheme for truncating the Hilbert space to only the most important determinants. We verify that this truncation introduces negligible error and use this new approach to simulate ASP for sets of small molecular systems and Hubbard models. Furthermore, we examine two approaches to speeding up ASP when performed on quantum hardware: (i) using the complete active space configuration interaction (CASCI) wave function instead of the Hartree–Fock initial state and (ii) a nonlinear interpolation between the initial and target Hamiltonians. We find that starting with a CASCI wave function with a limited active space yields substantial speedups for many of the systems examined, while nonlinear interpolation does not. In additional, we observe interesting trends in the minimum gap location (based on the initial state) as well as how state preparation time can depend on certain molecular properties, such as the number of valence electrons. Importantly, we find that the required state preparation times do not show an immediate exponential wall that would preclude an efficient run of ASP on actual hardware.

Kremenetski, Vladimir↗