SEARCH · Search NASA
Results for “accelerated 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.
Multi-fidelity classification using Gaussian processes: Accelerating the prediction of large-scale computational models
We report that machine learning techniques typically rely on large datasets to create accurate classifiers. However, there are situations when data is scarce and expensive to acquire. This is the case of studies that rely on state-of-the-art computational models which typically take days to run, thus hindering the potential of machine learning tools. In this work, we present a novel classifier that takes advantage of lower fidelity models and inexpensive approximations to predict the binary output of expensive computer simulations. We postulate an autoregressive model between the different levels of fidelity with Gaussian process priors. We adopt a fully Bayesian treatment for the hyper-parameters and use Markov Chain Monte Carlo samplers. We take advantage of the probabilistic nature of the classifier to implement active learning strategies. We also introduce a sparse approximation to enhance the ability of the multi-fidelity classifier to handle a large amount of low fidelity samples. We test these multi-fidelity classifiers against their single-fidelity counterpart with synthetic data, showing a median computational cost reduction of 23% for a target accuracy of 90%. In an application to cardiac electrophysiology, the multi-fidelity classifier achieves an F1 score, the harmonic mean of precision and recall, of 99.6% compared to 74.1% of a single-fidelity classifier when both are trained with 50 samples. In general, our results show that the multi-fidelity classifiers outperform their single-fidelity counterpart in terms of accuracy in all cases. Finally, we envision that this new tool will enable researchers to study classification problems that would otherwise be prohibitively expensive. Source code is available at https://github.com/fsahli/MFclass.
AMPS: An Augmented Matrix Formulation for Principal Submatrix Updates with Application to Power Grids
We present an augmented matrix approach to update the solution to a linear system of equations when the coefficient matrix is modified by a few elements within a principal submatrix. This problem arises in the dynamic security analysis of a power grid, where operators need to perform $N-x$ contingency analysis, i.e., determine the state of the system when up to $x$ links from $N$ fail. Our algorithms augment the coefficient matrix to account for the changes in it, and then compute the solution to the augmented system without refactoring the modified matrix. We provide two algorithms, a direct method, and a hybrid direct-iterative method for solving the augmented system. We also exploit the sparsity of the matrices and vectors to accelerate the overall computation. Our algorithms are compared on three power grids with PARDISO, a parallel direct solver, and CHOLMOD, a direct solver with the ability to modify the Cholesky factors of the coefficient matrix. We show that our augmented algorithms outperform PARDISO (by two orders of magnitude), and CHOLMOD (by a factor of up to 5). Further, our algorithms scale better than CHOLMOD as the number of elements updated increases. The solutions are computed with high accuracy. Our algorithms are capable of computing $N-x$ contingency analysis on a $778K$ bus grid, updating a solution with $x=20$ elements in $1.6 \times 10^{-2}$ seconds on an Intel Xeon processor.
GPU-Accelerated Lattice Boltzmann Method for Shock Physics Computations
Explore the source record for details and available documents.
Convergence acceleration and shock fitting for transonic aerodynamics computations
Two problems in computational fluid dynamics are studied in the context of transonic small-disturbance theory - namely, (1) how to speed up the convergence for currently available iterative procedures, and (2) how a shock-fitting method may be adapted to existing relaxation procedures with minimal alterations in computer programming and storage requirements. The paper contributes to a clarification of error analyses for sequence transformations based on the power method (including also the nonlinear transforms of Aitken, Shanks, and Wilkinson), and to developing a cyclic iterative procedure applying the transformations. Examples testing the procedure for a model Dirichlet problem and for a transonic airfoil problem show that savings in computer time by a factor of three to five are generally possible, depending on accuracy requirements and the particular iterative procedure used.-
Accelerating Incremental Checkpointing for Extreme-Scale Computing.
Abstract Not Provided
Accelerated pattern matching method on a quantum computing system
A method of determining a pattern in a sequence of bits using a quantum computing system includes setting a first register of a quantum processor in a superposition of a plurality of string index states, encoding a bit string in a second register of the quantum processor, encoding a bit pattern in a third register of the quantum processor, circularly shifting qubits of the second register conditioned on the first register, amplifying an amplitude of a state combined with the first register in which the circularly shifted qubits of the second register matches qubits of the third register, measuring an amplitude of the first register and determining a string index state of the plurality of string index states associated with the amplified state, and outputting, by use of a classical computer, a string index associated with the first register in the measured state.
Data-driven equation-free dynamics applied to many-protein complexes: The microtubule tip relaxation
Microtubules (MTs) constitute the largest components of the eukaryotic cytoskeleton and play crucial roles in various cellular processes, including mitosis and intracellular transport. The property allowing MTs to cater to such diverse roles is attributed to dynamic instability, which is coupled to the hydrolysis of GTP (guanosine-5'-triphosphate) to GDP (guanosine-5'-diphosphate) within the β-tubulin monomers. Understanding the equilibrium dynamics and the structural features of both GDP- and GTP-complexed MT tips, especially at an all-atom level, remains challenging for both experimental and computational methods because of their dynamic nature and the prohibitive computational demands of simulating large, many-protein systems. This study employs the “equation-free” multiscale computational method to accelerate the relaxation of all-atom simulations of MT tips toward their putative equilibrium conformation. Using large MT lattice systems (14 protofilaments × 8 heterodimers) comprising ~21-38 million atoms, we applied this multiscale approach to leapfrog through time and nearly double the computational efficiency in realizing relaxed all-atom conformations of GDP- and GTP-complexed MT tips. Commencing from an initial 4 μs unbiased all-atom simulation, we interleave coarse projective “equation-free” jumps with short bursts of all-atom molecular dynamics simulation to realize an additional effective simulation time of 1.875 μs. Our 5.875 μs of effective simulation trajectories for each system expose the subtle yet essential differences in the structures of MT tips as a function of whether β-tubulin monomer is complexed with GDP or GTP, as well as the lateral interactions within the MT tip, offering a refined understanding of features underlying MT dynamic instability. Furthermore, the approach presents a robust and generalizable framework for future explorations of large biomolecular systems at atomic resolution.
Quantum Noise Mitigation: Introducing the Robust Quantum Circuit Scheduler for Enhanced Fidelity and Throughput
Undoubtedly, quantum computing offers valuable acceleration for solving intricate problems. One of the primary hurdles lies in executing large-scale quantum applications on backend machines. Qubit noise, among other factors, dramatically influences the execution process. Implementing effective scheduling techniques for quantum circuits is crucial for practical quantum computing and preventing excessive waiting times. The quantum realm is distinct from classical computing in terms of optimization, performance, utilization, and waiting periods. Consequently, the parameters and components of quantum circuit scheduling diverge from those of classical computing. This paper presents Quantum Noise Mitigation: Introducing the Robust Quantum Circuit Scheduler for Enhanced Fidelity and Throughput, a straightforward yet effective scheduling framework and policy that enhances noise resilience, throughput, and the fidelity of quantum circuits. Drawing inspiration from classical methods, our scheduling approach incorporates additional constraints tailored for quantum logic. The outcome demonstrates a substantial improvement in fidelity and resource management, which is vital for real-world quantum applications.
Gravitational acceleration as a cue for absolute size and distance?
When an object's motion is influenced by gravity, as in the rise and fall of a thrown ball, the vertical component of acceleration is roughly constant at 9.8 m/sec2. In principle, an observer could use this information to estimate the absolute size and distance of the object (Saxberg, 1987a; Watson, Banks, von Hofsten, & Royden, 1992). In five experiments, we examined people's ability to utilize the size and distance information provided by gravitational acceleration. Observers viewed computer simulations of an object rising and falling on a trajectory aligned with the gravitational vector. The simulated objects were balls of different diameters presented across a wide range of simulated distances. Observers were asked to identify the ball that was presented and to estimate its distance. The results showed that observers were much more sensitive to average velocity than to the gravitational acceleration pattern. Likewise, verticality of the motion and visibility of the trajectory's apex had negligible effects on the accuracy of size and distance judgments.
Generating clock signals for a cycle accurate, cycle reproducible FPGA based hardware accelerator
A method, system and computer program product are disclosed for generating clock signals for a cycle accurate FPGA based hardware accelerator used to simulate operations of a device-under-test (DUT). In one embodiment, the DUT includes multiple device clocks generating multiple device clock signals at multiple frequencies and at a defined frequency ratio; and the FPG hardware accelerator includes multiple accelerator clocks generating multiple accelerator clock signals to operate the FPGA hardware accelerator to simulate the operations of the DUT. In one embodiment, operations of the DUT are mapped to the FPGA hardware accelerator, and the accelerator clock signals are generated at multiple frequencies and at the defined frequency ratio of the frequencies of the multiple device clocks, to maintain cycle accuracy between the DUT and the FPGA hardware accelerator. In an embodiment, the FPGA hardware accelerator may be used to control the frequencies of the multiple device clocks.
Accelerating the Design of Solar Thermal Fuel Materials through High Throughput Simulations
Solar thermal fuels (STF) store the energy of sunlight, which can then be released later in the form of heat, offering an emission-free and renewable solution for both solar energy conversion and storage. However, this approach is currently limited by the lack of low-cost materials with high energy density and high stability. In this Letter, we present an ab initio high-throughput computational approach to accelerate the design process and allow for searches over a broad class of materials. The high-throughput screening platform we have developed can run through large numbers of molecules composed of earth-abundant elements and identifies possible metastable structures of a given material. Corresponding isomerization enthalpies associated with the metastable structures are then computed. Using this high-throughput simulation approach, we have discovered molecular structures with high isomerization enthalpies that have the potential to be new candidates for high-energy density STF. We have also discovered physical principles to guide further STF materials design through structural analysis. More broadly, our results illustrate the potential of using high-throughput ab initio simulations to design materials that undergo targeted structural transitions.
Parameter Sweep and Optimization of Loosely Coupled Simulations Using the DAKOTA Toolkit
The increasing availability of large scale computing capabilities has accelerated the development of high-fidelity coupled simulations. Such simulations typically involve the integration of models that implement various aspects of the complex phenomena under investigation. Coupled simulations are playing an integral role in fields such as climate modeling, earth systems modeling, rocket simulations, computational chemistry, fusion research, and many other computational fields. Model coupling provides scientists with systematic ways to virtually explore the physical, mathematical, and computational aspects of the problem. Such exploration is rarely done using a single execution of a simulation, but rather by aggregating the results from many simulation runs that, together, serve to bring to light novel knowledge about the system under investigation. Furthermore, it is often the case (particularly in engineering disciplines) that the study of the underlying system takes the form of an optimization regime, where the control parameter space is explored to optimize an objective functions that captures system realizability, cost, performance, or a combination thereof. Novel and flexible frameworks that facilitate the integration of the disparate models into a holistic simulation are used to perform this research, while making efficient use of the available computational resources. In this paper, we describe the integration of the DAKOTA optimization and parameter sweep toolkit with the Integrated Plasma Simulator (IPS), a component-based framework for loosely coupled simulations. The integration allows DAKOTA to exploit the internal task and resource management of the IPS to dynamically instantiate simulation instances within a single IPS instance, allowing for greater control over the trade-off between efficiency of resource utilization and time to completion. We present a case study showing the use of the combined DAKOTA-IPS system to aid in the design of a lithium ion battery (LIB) cell, by studying a coupled system involving the electrochemistry and ion transport at the lower length scales and thermal energy transport at the device scales. The DAKOTA-IPS system provides a flexible tool for use in optimization and parameter sweep studies involving loosely coupled simulations that is suitable for use in situations where changes to the constituent components in the coupled simulation are impractical due to intellectual property or code heritage issues.
Economic impact in Illinois
As the United States’ premier particle physics and accelerator laboratory, Fermilab seeks to solve the mysteries of our universe. Located in Batavia, Illinois, the lab designs and builds the world’s most advanced particle accelerators, detectors and computing facilities. In 2018, the total annual economic impact in Illinois came to $452 million. New construction and research projects will increase that number in the coming years.
Finding predictive models for singlet fission by machine learning
Singlet fission (SF), the conversion of one singlet exciton into two triplet excitons, could significantly enhance solar cell efficiency. Molecular crystals that undergo SF are scarce. Computational exploration may accelerate the discovery of SF materials. However, many-body perturbation theory (MBPT) calculations of the excitonic properties of molecular crystals are impractical for large-scale materials screening. We use the sure-independence-screening-and-sparsifying-operator (SISSO) machine-learning algorithm to generate computationally efficient models that can predict the MBPT thermodynamic driving force for SF for a dataset of 101 polycyclic aromatic hydrocarbons (PAH101). SISSO generates models by iteratively combining physical primary features. The best models are selected by linear regression with cross-validation. The SISSO models successfully predict the SF driving force with errors below 0.2 eV. Based on the cost, accuracy, and classification performance of SISSO models, we propose a hierarchical materials screening workflow. Three potential SF candidates are found in the PAH101 set.
Performance of a geometric deep learning pipeline for HL-LHC particle tracking
The Exa.TrkX project has applied geometric learning concepts such as metric learning and graph neural networks to HEP particle tracking. Exa.TrkX’s tracking pipeline groups detector measurements to form track candidates and filters them. The pipeline, originally developed using the TrackML dataset (a simulation of an LHC-inspired tracking detector), has been demonstrated on other detectors, including DUNE Liquid Argon TPC and CMS High-Granularity Calorimeter. This paper documents new developments needed to study the physics and computing performance of the Exa.TrkX pipeline on the full TrackML dataset, a first step towards validating the pipeline using ATLAS and CMS data. The pipeline achieves tracking efficiency and purity similar to production tracking algorithms. Crucially for future HEP applications, the pipeline benefits significantly from GPU acceleration, and its computational requirements scale close to linearly with the number of particles in the event.
PABO: Pseudo Agent-Based Multi-Objective Bayesian Hyperparameter Optimization for Efficient Neural Accelerator Design
The ever increasing computational cost of Deep Neural Networks (DNN) and the demand for energy efficient hardware for DNN acceleration has made accuracy and hardware cost co-optimization for DNNs tremendously important, especially for edge devices. Owing to the large parameter space and cost of evaluating each parameter in the search space, manually tuning of DNN hyperparameters is impractical. Automatic joint DNN and hardware hyperparameter optimization is indispensable for such problems. Bayesian optimization-based approaches have shown promising results for hyperparameter optimization of DNNs. However, most of these techniques have been developed without considering the underlying hardware, thereby leading to inefficient designs. Further, the few works that perform joint optimization are not generalizable and mainly focus on CMOS-based architectures. In this work, we present a novel pseudo agent-based multiobjective hyperparameter optimization (PABO) for maximizing the DNN performance while obtaining low hardware cost. Compared to the existing methods, our work poses a theoretically different approach for joint optimization of accuracy and hardware cost and focuses on memristive crossbar based accelerators. PABO uses a supervisor agent to establish connections between the posterior Gaussian distribution models of network accuracy and hardware cost requirements. The agent reduces the mathematical complexity of the co-optimization problem by removing unnecessary computations and updates of acquisition functions, thereby achieving significant speed-ups for the optimization procedure. PABO outputs a Pareto frontier that underscores the trade-offs between designing high-accuracy and hardware efficiency. Our results demonstrate a superior performance compared to the state-of-the-art methods both in terms of accuracy and computational speed (~100x speed up).
Detector Simulation Challenges for Future Accelerator Experiments
Detector simulation is a key component for studies on prospective future high-energy colliders, the design, optimization, testing and operation of particle physics experiments, and the analysis of the data collected to perform physics measurements. This review starts from the current state of the art technology applied to detector simulation in high-energy physics and elaborates on the evolution of software tools developed to address the challenges posed by future accelerator programs beyond the HL-LHC era, into the 2030–2050 period. New accelerator, detector, and computing technologies set the stage for an exercise in how detector simulation will serve the needs of the high-energy physics programs of the mid 21st century, and its potential impact on other research domains.