SEARCH · Search NASA

Results for “Memory Optimization”

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.

SourceResearch connectionFrom yearThrough year

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

At least 55 records · Page 3

All sources NASA NTRS DOE OSTI

DOE OSTI2022

GPU-accelerated DNS of compressible turbulent flows

Here, this paper explores strategies to transform an existing CPU-based high-performance computational fluid dynamics solver, HyPar, for compressible flow simulations on emerging exascale heterogeneous (CPU+GPU) computing platforms. The scientific motivation for developing a GPU-enhanced version of HyPar is to simulate canonical turbulent flows at the highest resolution possible on such platforms. We show that optimizing memory operations and thread blocks results in 200x speedup of computationally intensive kernels compared with a CPU core. Using multiple GPUs and CUDA-aware MPI communication, we demonstrate both strong and weak scaling of our GPU-based HyPar implementation on the NVIDIA Volta V100 GPUs. We simulate the decay of homogeneous isotropic turbulence in a triply periodic box on grids with up to 1024 3 points (5.3 billion degrees of freedom) and on up to 1,024 GPUs. We compare the wall times for CPU-only and CPU+GPU simulations. The results presented in the paper are obtained on the Summit and Lassen supercomputers at Oak Ridge and Lawrence Livermore National Laboratories, respectively.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2024

Asynchronous GPU-based DEM solver embedded in commercial CFD software with polyhedral mesh support

A novel graphical processing unit-based discrete element method solver is introduced to improve stability, performance, and provide seamless integration into commercial or open-source computational fluid dynamics software. A key innovation is eliminating a need for network communication between solvers, which was previously required for cross-platform coupling. This is accomplished by a direct coupling method that employs dynamic-linked libraries. Furthermore, the solver optimizes memory usage by streamlining the particle-cell search algorithm by eliminating the cells' searching grid. This ensures the solver is compatible with a wide range of mesh types, providing high geometric flexibility. The approach simplifies the simulation process by directly incorporating computational fluid dynamics mesh information into the discrete element method solver. The performance analysis indicates about sixteen times boost in computational speed compared to benchmark central processing unit-based solvers. Finally, the solver's compatibility with polyhedral meshes, a vital advantage for complex geometries, is tested against a referenced study regarding the simulation of an immersed-tube fluidized bed.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

DOE OSTI2022

Automatic Differentiation of C++ Codes on Emerging Manycore Architectures with Sacado

Automatic differentiation (AD) is a well-known technique for evaluating analytic derivatives of calculations implemented on a computer, with numerous software tools available for incorporating AD technology into complex applications. However, a growing challenge for AD is the efficient differentiation of parallel computations implemented on emerging manycore computing architectures such as multicore CPUs, GPUs, and accelerators as these devices become more pervasive. In this work, we explore forward mode, operator overloading-based differentiation of C++ codes on these architectures using the widely available Sacado AD software package. In particular, we leverage Kokkos, a C++ tool providing APIs for implementing parallel computations that is portable to a wide variety of emerging architectures. Here we describe the challenges that arise when differentiating code for these architectures using Kokkos, and two approaches for overcoming them that ensure optimal memory access patterns as well as expose additional dimensions of fine-grained parallelism in the derivative calculation. We describe the results of several computational experiments that demonstrate the performance of the approach on a few contemporary CPU and GPU architectures. We then conclude with applications of these techniques to the simulation of discretized systems of partial differential equations.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2013

GPU-Accelerated Denoising in 3D (GD3D)

The raw computational power GPU Accelerators enables fast denoising of 3D MR images using bilateral filtering, anisotropic diffusion, and non-local means. This software addresses two facets of this promising application: what tuning is necessary to achieve optimal performance on a modern GPU? And what parameters yield the best denoising results in practice? To answer the first question, the software performs an autotuning step to empirically determine optimal memory blocking on the GPU. To answer the second, it performs a sweep of algorithm parameters to determine the combination that best reduces the mean squared error relative to a noiseless reference image.

Howison, Mark↗

DOE OSTI2023

End-to-end GPU acceleration of low-order-refined preconditioning for high-order finite element discretizations

In this article, we present algorithms and implementations for the end-to-end GPU acceleration of matrix-free low-order-refined preconditioning of high-order finite element problems. The methods described here allow for the construction of effective preconditioners for high-order problems with optimal memory usage and computational complexity. The preconditioners are based on the construction of a spectrally equivalent low-order discretization on a refined mesh, which is then amenable to, for example, algebraic multigrid preconditioning. The constants of equivalence are independent of mesh size and polynomial degree. For vector finite element problems in H(curl) and H(div) (e.g., for electromagnetic or radiation diffusion problems), a specially constructed interpolation–histopolation basis is used to ensure fast convergence. Detailed performance studies are carried out to analyze the efficiency of the GPU algorithms. The kernel throughput of each of the main algorithmic components is measured, and the strong and weak parallel scalability of the methods is demonstrated. The different relative weighting and significance of the algorithmic components on GPUs and CPUs is discussed. Results on problems involving adaptively refined nonconforming meshes are shown, and the use of the preconditioners on a large-scale magnetic diffusion problem using all spaces of the finite element de Rham complex is illustrated.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2022

SMALE: Enhancing Scalability of Machine Learning Algorithms on Extreme-Scale Computing Platforms

Deployment and execution of machine learning tasks on extreme-scale computing platforms face several significant technical challenges: 1) High computing cost incurred by dense networks – The computing workload of deep networks with densely-connected topology increases rapidly with the network size, imposing a non-scalable computing model of extreme-scale computing platforms; 2) Non-optimized workload distribution – Many advanced deep learning algorithms, e.g., sparsification and irregular net-work topology, produce very unbalanced workload distribution on extreme-scale computing platforms. The computation efficiency is greatly hindered by the incurred data and computation redundancies as well as long tails of the node with extensive workload; 3) Constraints in data movement and I/O bottle-neck – Inter-node data movement in extreme-scale computing platforms are associated with high energy and latency costs, and subject to the constraints of I/O bandwidth; and 4) Generalization of algorithm realization and acceleration on computing platforms – The large varieties of machine learning algorithms and structures of extreme-scale computing platforms make the derivation of a generalized algorithm realization and acceleration method very challenging, which, however, is the requirement by domain scientists and interested users. We call the above challenges Smale’s Problems in Machine Learning and Understanding for High-Performance Computing Scientific Discovery. The objective of our three-year research project is to develop a holistic innovation set at structure, assembly, and acceleration layers of machine learning algorithms to address the above challenges in algorithm deployment and execution. Three tasks are particularly performed, including: At the algorithm structure level, we investigate the techniques that can structurally sparsify on the topology of deep networks for computing workload reduction. We also study clustering and pruning techniques that can optimize the workload distributions over the extreme-scale computing platforms; At the algorithm assembly level, we derive a unified learning framework for unsupervised transfer learning and dynamic growing capabilities. Novel training methods are also exploited to enhance the training efficiency of the proposed framework; At the algorithm acceleration level, we will develop a series of techniques that can accelerate the computation of sparse matrix operations, which are one of the core executions in deep learning and optimize memory access of the concerned platforms. Our proposed techniques attack the fundamental problems in machine learning algorithms running on extreme-scale computing platforms by vertically integrating the solutions at three closely entangled layers, paving the long-term scaling path of machine learning applications under DOE context. Three tasks corresponding to the above respective research orientations are performed during the three-year project period with our collaborators at ORNL. The outcome of the proposed project is anticipated to form a holistic solution set of novel algorithms and network topologies, efficient training techniques, and fast acceleration methods to promote the computing scalability of the machine learning applications of particular interest to DOE.

97 MATHEMATICS AND COMPUTING↗

NASA NTRS1993

Parallel algorithm for dominant points correspondences in robot binocular stereo vision

This paper presents an algorithm to find the correspondences of points representing dominant feature in robot stereo vision. The algorithm consists of two main steps: dominant point extraction and dominant point matching. In the feature extraction phase, the algorithm utilizes the widely used Moravec Interest Operator and two other operators: the Prewitt Operator and a new operator called Gradient Angle Variance Operator. The Interest Operator in the Moravec algorithm was used to exclude featureless areas and simple edges which are oriented in the vertical, horizontal, and two diagonals. It was incorrectly detecting points on edges which are not on the four main directions (vertical, horizontal, and two diagonals). The new algorithm uses the Prewitt operator to exclude featureless areas, so that the Interest Operator is applied only on the edges to exclude simple edges and to leave interesting points. This modification speeds-up the extraction process by approximately 5 times. The Gradient Angle Variance (GAV), an operator which calculates the variance of the gradient angle in a window around the point under concern, is then applied on the interesting points to exclude the redundant ones and leave the actual dominant ones. The matching phase is performed after the extraction of the dominant points in both stereo images. The matching starts with dominant points in the left image and does a local search, looking for corresponding dominant points in the right image. The search is geometrically constrained the epipolar line of the parallel-axes stereo geometry and the maximum disparity of the application environment. If one dominant point in the right image lies in the search areas, then it is the corresponding point of the reference dominant point in the left image. A parameter provided by the GAV is thresholded and used as a rough similarity measure to select the corresponding dominant point if there is more than one point the search area. The correlation is used as a final decision tool when there is still more than one point in the search area. If there is no dominant point in the search area of if the points in the search area are below a correlation threshold, then the dominant point in the reference image is occluded and can not be corresponded. The algorithm has been modeled, implemented and shown to be fast, robust and parallel. The parallelism is created from three main features: locality of the operators; a memory optimization scheme; and the ability to fully parallelize the extraction phase which is the most computational intensive task in the algorithm. The last feature is achieved by performing the extraction phase on the two images simultaneously.

Al-Tammami, A.↗

NASA NTRS2006

Architecture Adaptive Computing Environment

Architecture Adaptive Computing Environment (aCe) is a software system that includes a language, compiler, and run-time library for parallel computing. aCe was developed to enable programmers to write programs, more easily than was previously possible, for a variety of parallel computing architectures. Heretofore, it has been perceived to be difficult to write parallel programs for parallel computers and more difficult to port the programs to different parallel computing architectures. In contrast, aCe is supportable on all high-performance computing architectures. Currently, it is supported on LINUX clusters. aCe uses parallel programming constructs that facilitate writing of parallel programs. Such constructs were used in single-instruction/multiple-data (SIMD) programming languages of the 1980s, including Parallel Pascal, Parallel Forth, C*, *LISP, and MasPar MPL. In aCe, these constructs are extended and implemented for both SIMD and multiple- instruction/multiple-data (MIMD) architectures. Two new constructs incorporated in aCe are those of (1) scalar and virtual variables and (2) pre-computed paths. The scalar-and-virtual-variables construct increases flexibility in optimizing memory utilization in various architectures. The pre-computed-paths construct enables the compiler to pre-compute part of a communication operation once, rather than computing it every time the communication operation is performed.

Dorband, John E.↗

DOE OSTI2023

IP Protection in TinyML

Tiny machine learning (TinyML) is an essential component of emerging smart microcontrollers (MCUs). How- ever, the protection of the intellectual property (IP) of the model is an increasing concern due to the lack of desktop/server-grade resources on these power-constrained devices. In this paper, we propose STML, a system and algorithm co-design to Secure IP of TinyML on MCUs with ARM TrustZone. Our design jointly optimizes memory utilization and latency while ensuring the security and accuracy of emerging models. We implemented a prototype and benchmarked with 7 models, demonstrating STML reduces 40% of model protection runtime overhead on average.

42 ENGINEERING↗

DOE OSTI2024

Commissioning of the Mu2e tracker DAQ, planning for the Vertical Slice Test and pre-pattern recognition studies

The primary objective of the Mu2e experiment at Fermilab is to search for the neutrino-less coherent $\mu \rightarrow e$ conversion in the field of an aluminum nucleus ($\mu^- \text{Al} \rightarrow e^- \text{Al}$). The signature of this process is a monochromatic Conversion Electron (CE) with an energy of approximately 104.97 MeV \cite{bartoszek2015mu2e}. Within the Standard Model (SM), the branching ratio for this process, including neutrino masses and oscillation, is expected to be less than $\mathcal{O}(10^{-50})$. This value is far beyond current experimental capabilities. However, models of physics beyond the SM predict much higher relative rates, approaching an observable level. The SINDRUM II experiment set an upper limit on muon conversion at $7 \times 10^{-13}$ (90\% CL) on Au target \cite{SINDRUMII:2006dvw}, and the Mu2e collaboration aims to improve this limit by four orders of magnitude. Observing this process would provide a clear evidence of physics beyond the Standard Model. A brief discussion of the theoretical and experimental aspects is provided in Chapter \ref{intr}. Mu2e adopts a sophisticated experimental setup to achieve its goals, further described in Chapter \ref{mu2echapter}. The central part of the Mu2e detector is the tracker, that consists of 18 tracking stations. The tracker must provide excellent momentum resolution, approximately 1 MeV/c, to distinguish the monochromatic CE signal from the background. To minimize the energy losses, a straw tube tracker will be used \cite{bobbb}. Chapter \ref{chaptertrk} provides an overview of the straw tracker design and its working principles. This Thesis presents a comprehensive study of the Mu2e tracker, covering complementary aspects from initial commissioning to optimization and first steps of the calibration processes. My work at Fermilab has been focused on the complete Data Acquisition (DAQ) testing from both hardware and software perspectives. I was involved in the commissioning of the Mu2e DAQ system and the Vertical Slice Test (VST) of the tracker. The VST encompasses the entire testing chain, from the straws to the readout, and to processed data on disk. I was also focused on the offline analysis, especially on pre-pattern recognition studies, to explore the best methods for identifying $\delta$-electrons during the data taking. Chapter \ref{commissioning} details the commissioning of the tracker DAQ system, emphasizing the importance of understanding of the readout process before the data acquisition. This includes validating the readout logic and firmware through Monte Carlo simulations to confirm functionality and buffering, monitoring the quality of the data from the tracker preamplifiers and front-end electronics, and assessing overall DAQ performance to ensure reliability during future calibration and data-taking. Chapter \ref{planning} discusses the initial steps towards the tracker calibration. The ultimate goal is to perform a time calibration of the first assembled station of the tracker using cosmic muons, aiming for a longitudinal hit position resolution better than 4 cm. This involves determining the signal propagation times and channel-to-channel delays. I performed a Monte Carlo study to determine the impact of the station orientation on the quality of the calibration, in particular on the cosmic track reconstruction, focusing on potential biases that could arise. These studies provide essential insights into the operation, optimization, and calibration of the Mu2e tracker system. Given the high data volume expected during Mu2e operations, estimated at approximately 7 PBytes per year, optimizing memory usage and minimizing CPU consumption are critical. A significant challenge lies in effectively flagging $\delta$-electron hits, which are the primary source of hits in the tracker, without compromising the efficiency of CE hit detection and track reconstruction. A detailed study of pre-pattern recognition and a thorough comparison of two $\delta$-electron flagging algorithms is provided in Chapter \ref{delta}. In Chapter \ref{conclusions}, the findings are concisely summarized, offering a comprehensive synthesis of the research and emphasizing the key insights derived from this study.

43 PARTICLE ACCELERATORS↗

NASA NTRS1979

Some optimal partial-unit-memory codes

A class of time-invariant binary convolutional codes is defined, called partial-unit-memory codes. These codes are optimal in the sense of having maximum free distance for given values of R, k (the number of encoder inputs), and mu (the number of encoder memory cells). Optimal codes are given for rates R = 1/4, 1/3, 1/2, and 2/3, with mu not greater than 4 and k not greater than mu + 3, whenever such a code is better than previously known codes. An infinite class of optimal partial-unit-memory codes is also constructed based on equidistant block codes.

Lauer, G. S.↗

DOE OSTI2017

Recall Performance for Content-Addressable Memory Using Adiabatic Quantum Optimization

A content-addressable memory (CAM) stores key-value associations such that the key is recalled by providing its associated value. While CAM recall is traditionally performed using recurrent neural network models, we show how to solve this problem using adiabatic quantum optimization. Our approach maps the recurrent neural network to a commercially available quantum processing unit by taking advantage of the common underlying Ising spin model. We then assess the accuracy of the quantum processor to store key-value associations by quantifying recall performance against an ensemble of problem sets. We observe that different learning rules from the neural network community influence recall accuracy but performance appears to be limited by potential noise in the processor. The strong connection established between quantum processors and neural network problems supports the growing intersection of these two ideas.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2014

Adiabatic quantum optimization for associative memory recall

Hopfield networks are a variant of associative memory that recall patterns stored in the couplings of an Ising model. Stored memories are conventionally accessed as fixed points in the network dynamics that correspond to energetic minima of the spin state. We show that memories stored in a Hopfield network may also be recalled by energy minimization using adiabatic quantum optimization (AQO). Numerical simulations of the underlying quantum dynamics allow us to quantify AQO recall accuracy with respect to the number of stored memories and noise in the input key. We investigate AQO performance with respect to how memories are stored in the Ising model according to different learning rules. Our results demonstrate that AQO recall accuracy varies strongly with learning rule, a behavior that is attributed to differences in energy landscapes. Consequently, learning rules offer a family of methods for programming adiabatic quantum optimization that we expect to be useful for characterizing AQO performance.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

DOE OSTI2016

Strategic priming with multiple antigens can yield memory cell phenotypes optimized for infection with Mycobacterium tuberculosis : A computational study

Lack of an effective vaccine results in 9 million new cases of tuberculosis (TB) every year and 1.8 million deaths worldwide. While many infants are vaccinated at birth with BCG (an attenuated M. bovis ), this does not prevent infection or development of TB after childhood. Immune responses necessary for prevention of infection or disease are still unknown, making development of effective vaccines against TB challenging. Several new vaccines are ready for human clinical trials, but these trials are difficult and expensive; especially challenging is determining the appropriate cellular response necessary for protection. The magnitude of an immune response is likely key to generating a successful vaccine. Characteristics such as numbers of central memory (CM) and effector memory (EM) T cells responsive to a diverse set of epitopes are also correlated with protection. Promising vaccines against TB contain mycobacterial subunit antigens (Ag) present during both active and latent infection. We hypothesize that protection against different key immunodominant antigens could require a vaccine that produces different levels of EM and CM for each Ag-specific memory population. We created a computational model to explore EM and CM values, and their ratio, within what we term Memory Design Space. Our model captures events involved in T cell priming within lymph nodes and tracks their circulation through blood to peripheral tissues. We used the model to test whether multiple Ag-specific memory cell populations could be generated with distinct locations within Memory Design Space at a specific time point post vaccination. Boosting can further shift memory populations to memory cell ratios unreachable by initial priming events. By strategically varying antigen load, properties of cellular interactions within the LN, and delivery parameters (e.g., number of boosts) of multi-subunit vaccines, we can generate multiple Ag-specific memory populations that cover a wide range of Memory Design Space. As a result, given a set of desired characteristics for Ag-specific memory populations, we can use our model as a tool to predict vaccine formulations that will generate those populations.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2018

Data Placement Optimization in GPU Memory Hierarchy using Predictive Modeling

Explore the source record for details and available documents.

Computer science↗

NASA NTRSDate not supplied

Optimized Entanglement Distribution Among Quantum Memories

We evaluate entanglement distribution performance when an optimized quantum memory encoding method is used for two sources, the standard spontaneous parametric down conversion source and the zero added loss multiplexing source. Physical non-idealities are considered.

Entanglement distribution↗

NASA NTRS1990

Execution time supports for adaptive scientific algorithms on distributed memory machines

Optimizations are considered that are required for efficient execution of code segments that consists of loops over distributed data structures. The PARTI (Parallel Automated Runtime Toolkit at ICASE) execution time primitives are designed to carry out these optimizations and can be used to implement a wide range of scientific algorithms on distributed memory machines. These primitives allow the user to control array mappings in a way that gives an appearance of shared memory. Computations can be based on a global index set. Primitives are used to carry out gather and scatter operations on distributed arrays. Communications patterns are derived at runtime, and the appropriate send and receive messages are automatically generated.

Berryman, Harry↗

NASA NTRS1990

Execution time support for scientific programs on distributed memory machines

Berryman, Harry↗