Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “limited memory”

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 37 records · Page 2

L-BFGS Class Implementation in C++

This report presents a header-only C++ class implementation of the Limited-memory BroydenFletcher-Goldfarb-Shanno (L-BFGS) algorithm. The L-BFGS method is a general purpose quasi-Netwon optimization method that builds an approximation of the descent direction from consecutive iterate and gradient vectors. The limited-memory aspect stems from the replacement of the N × N approximation matrix of the original BFGS method with M vectors of length N. An example usage of the class is included along with the reference source code.

97 MATHEMATICS AND COMPUTING↗

A Topographical Lidar System for Terrain-Relative Navigation

An imaging lidar system is being developed for use in navigation, relative to the local terrain. This technology will potentially be used for future spacecraft landing on the Moon. Systems like this one could also be used on Earth for diverse purposes, including mapping terrain, navigating aircraft with respect to terrain and military applications. The system has been field-tested aboard a helicopter in the Mojave Desert. When this system was designed, digitizers with sufficient sampling rate (2 GHz) were only available with very limited memory. Also, it was desirable to limit the amount of data to be transferred between the digitizer and the mass storage between individual frames. One of the novelty design features of this system was to design the system around the limited amount of memory of the digitizer. The system is required to operate over an altitude (distance) range from a few meters to approximately 1 km, but for each scan across the full field of view, the digitizer memory is only able to hold data for an altitude range no more than 100 m. Data acquisition methods in support of the limited 100 m wide altitude range are described.

Liebe, Carl Christian↗

SU-E-T-395: Multi-GPU-Based VMAT Treatment Plan Optimization Using a Column-Generation Approach

Purpose: GPU has been employed to speed up VMAT optimizations from hours to minutes. However, its limited memory capacity makes it difficult to handle cases with a huge dose-deposition-coefficient (DDC) matrix, e.g. those with a large target size, multiple arcs, small beam angle intervals and/or small beamlet size. We propose multi-GPU-based VMAT optimization to solve this memory issue to make GPU-based VMAT more practical for clinical use. Methods: Our column-generation-based method generates apertures sequentially by iteratively searching for an optimal feasible aperture (referred as pricing problem, PP) and optimizing aperture intensities (referred as master problem, MP). The PP requires access to the large DDC matrix, which is implemented on a multi-GPU system. Each GPU stores a DDC sub-matrix corresponding to one fraction of beam angles and is only responsible for calculation related to those angles. Broadcast and parallel reduction schemes are adopted for inter-GPU data transfer. MP is a relatively small-scale problem and is implemented on one GPU. One headand- neck cancer case was used for test. Three different strategies for VMAT optimization on single GPU were also implemented for comparison: (S1) truncating DDC matrix to ignore its small value entries for optimization; (S2) transferring DDC matrix part by part to GPU during optimizations whenever needed; (S3) moving DDC matrix related calculation onto CPU. Results: Our multi-GPU-based implementation reaches a good plan within 1 minute. Although S1 was 10 seconds faster than our method, the obtained plan quality is worse. Both S2 and S3 handle the full DDC matrix and hence yield the same plan as in our method. However, the computation time is longer, namely 4 minutes and 30 minutes, respectively. Conclusion: Our multi-GPU-based VMAT optimization can effectively solve the limited memory issue with good plan quality and high efficiency, making GPUbased ultra-fast VMAT planning practical for real clinical use.

60 APPLIED LIFE SCIENCES↗

Efficiently modeling neural networks on massively parallel computers

Neural networks are a very useful tool for analyzing and modeling complex real world systems. Applying neural network simulations to real world problems generally involves large amounts of data and massive amounts of computation. To efficiently handle the computational requirements of large problems, we have implemented at Los Alamos a highly efficient neural network compiler for serial computers, vector computers, vector parallel computers, and fine grain SIMD computers such as the CM-2 connection machine. This paper describes the mapping used by the compiler to implement feed-forward backpropagation neural networks for a SIMD (Single Instruction Multiple Data) architecture parallel computer. Thinking Machines Corporation has benchmarked our code at 1.3 billion interconnects per second (approximately 3 gigaflops) on a 64,000 processor CM-2 connection machine (Singer 1990). This mapping is applicable to other SIMD computers and can be implemented on MIMD computers such as the CM-5 connection machine. Our mapping has virtually no communications overhead with the exception of the communications required for a global summation across the processors (which has a sub-linear runtime growth on the order of O(log(number of processors)). We can efficiently model very large neural networks which have many neurons and interconnects and our mapping can extend to arbitrarily large networks (within memory limitations) by merging the memory space of separate processors with fast adjacent processor interprocessor communications. This paper will consider the simulation of only feed forward neural network although this method is extendable to recurrent networks.

Farber, Robert M.↗

Parallelization and checkpointing of GPU applications through program transformation

GPUs have emerged as a powerful tool for accelerating general-purpose applications. The availability of programming languages that makes writing general-purpose applications for running on GPUs tractable have consolidated GPUs as an alternative for accelerating general purpose applications. Among the areas that have benefited from GPU acceleration are: signal and image processing, computational fluid dynamics, quantum chemistry, and, in general, the High Performance Computing (HPC) Industry. In order to continue to exploit higher levels of parallelism with GPUs, multi-GPU systems are gaining popularity. In this context, single-GPU applications are parallelized for running in multi-GPU systems. Furthermore, multi-GPU systems help to solve the GPU memory limitation for applications with large application memory footprint. Parallelizing single-GPU applications has been approached by libraries that distribute the workload at runtime, however, they impose execution overhead and are not portable. On the other hand, on traditional CPU systems, parallelization has been approached through application transformation at pre-compile time, which enhances the application to distribute the workload at application level and does not have the issues of library-based approaches. Hence, a parallelization scheme for GPU systems based on application transformation is needed. Like any computing engine of today, reliability is also a concern in GPUs. GPUs are vulnerable to transient and permanent failures. Current checkpoint/restart techniques are not suitable for systems with GPUs. Checkpointing for GPU systems present new and interesting challenges, primarily due to the natural differences imposed by the hardware design, the memory subsystem architecture, the massive number of threads, and the limited amount of synchronization among threads. Therefore, a checkpoint/restart technique suitable for GPU systems is needed. The goal of this work is to exploit higher levels of parallelism and to develop support for application-level fault tolerance in applications using multiple GPUs. Our techniques reduce the burden of enhancing single-GPU applications to support these features. To achieve our goal, this work designs and implements a framework for enhancing a single-GPU OpenCL application through application transformation.

97 MATHEMATICS AND COMPUTING↗

Randomized Sketching Algorithms for Low-Memory Dynamic Optimization

This paper develops a novel limited-memory method to solve dynamic optimization problems. The memory requirements for such problems often present a major obstacle, particularly for problems with PDE constraints such as optimal flow control, full waveform inversion, and optical tomography. In these problems, PDE constraints uniquely determine the state of a physical system for a given control; the goal is to find the value of the control that minimizes an objective. While the control is often low dimensional, the state is typically more expensive to store. This paper suggests using randomized matrix approximation to compress the state as it is generated and shows how to use the compressed state to reliably solve the original dynamic optimization problem. Concretely, the compressed state is used to compute approximate gradients and to apply the Hessian to vectors. The approximation error in these quantities is controlled by the target rank of the sketch. This approximate first- and second-order information can readily be used in any optimization algorithm. As an example, we develop a sketched trust-region method that adaptively chooses the target rank using a posteriori error information and provably converges to a stationary point of the original problem. Numerical experiments with the sketched trust-region method show promising performance on challenging problems such as the optimal control of an advection-reaction-diffusion equation and the optimal control of fluid flow past a cylinder.

97 MATHEMATICS AND COMPUTING↗

Large-Scale Optimization with Linear Equality Constraints Using Reduced Compact Representation

For optimization problems with linear equality constraints, we prove that the (1,1) block of the inverse KKT matrix remains unchanged when projected onto the nullspace of the constraint matrix. In this work, we develop reduced compact representations of the limited-memory inverse BFGS Hessian to compute search directions efficiently when the constraint Jacobian is sparse. Orthogonal projections are implemented by a sparse QR factorization or a preconditioned LSQR iteration. In numerical experiments two proposed trust-region algorithms improve in computation times, often significantly, compared to previous implementations of related algorithms and compared to IPOPT.

97 MATHEMATICS AND COMPUTING↗

An adaptive Hessian approximated stochastic gradient MCMC method

Bayesian approaches have been successfully integrated into training deep neural networks. One popular family is stochastic gradient Markov chain Monte Carlo methods (SG-MCMC), which have gained increasing interest due to their ability to handle large datasets and the potential to avoid overfitting. Although standard SG-MCMC methods have shown great performance in a variety of problems, they may be inefficient when the random variables in the target posterior densities have scale differences or are highly correlated. Here, we present an adaptive Hessian approximated stochastic gradient MCMC method to incorporate local geometric information while sampling from the posterior. The idea is to apply stochastic approximation (SA) to sequentially update a preconditioning matrix at each iteration. The preconditioner possesses second-order information and can guide the random walk of a sampler efficiently. Instead of computing and saving the full Hessian of the log posterior, we use limited memory of the samples and their stochastic gradients to approximate the inverse Hessian-vector multiplication in the updating formula. Moreover, by smoothly optimizing the preconditioning matrix via SA, our proposed algorithm can asymptotically converge to the target distribution with a controllable bias under mild conditions. To reduce the training and testing computational burden, we adopt a magnitude-based weight pruning method to enforce the sparsity of the network. Our method is user-friendly and demonstrates better learning results compared to standard SG-MCMC updating rules. The approximation of inverse Hessian alleviates storage and computational complexities for large dimensional models. Numerical experiments are performed on several problems, including sampling from 2D correlated distribution, synthetic regression problems, and learning the numerical solutions of heterogeneous elliptic PDE. The numerical results demonstrate great improvement in both the convergence rate and accuracy.

97 MATHEMATICS AND COMPUTING↗

Graph-based Reversible Evaluation and Tangents Library

GRETL is a C++ library for evaluation, re-evaluation and algorithmic differentiation of functional operations on an arbitrary computational graph with limited memory usage. Similar to popular machine learning frameworks in Python, like PyTorch and JAX, it tracks and stores both operations and output data as functions are evaluated. Once this composition of functions is built up, the entire chain of operations can be back propagated to compute sensitivities of the final result with respect to any number of inputs. In contrast to most machine learning applications, memory usage becomes the bottleneck for back propagation in many physics applications, especially for time-dependent PDEs. Dynamic check pointing becomes essential. An important distinguishing feature of GRETL is its ability to limit the maximum memory usage by automatically dynamic checkpointing the data output for each graph operation (see Wang, Moin, Iaccarino, 2009). During backpropagation, parts of the graph that are no longer in memory are automatically re-evaluated from upstream checkpointed states as needed for derivative sensitivity calculations (or more precisely, for vector-Jacobian products). GRETL is particularly beneficial for applications, such as coupled multi-physics, where deriving adjoint-based sensitivities and managing checkpoint memory across modules becomes onerous. Cases which can be readily handled by the GRETL library include: different time-integration algorithms per physics (e.g., coupled predictor-corrector algorithms, IMEX, etc.), sub-cycling, asynchronous integrators, state dependent timestep sizes, iterative solvers and coupling algorithms, controller algorithms, and more.

Tupek, MichaelR [Lawrence Livermore National Labor↗

Umpire: Status Report and Future Development Plan

Umpire is a open-source C++ library that provides memory and resource management functions for use with heterogeneous memory spaces. The Umpire library is being developed to simplify coordination and sharing of limited memory resources for complex integrated applications developed in the ECP/ATDM and ASC programs, and other multi-physics codes that combine many physics packages and libraries.

97 MATHEMATICS AND COMPUTING↗

Umpire: Integration with Other Software Technologies

Umpire is an open-source C++ library that provides memory and resource management functions for use with heterogeneous memory spaces. The Umpire library is being developed to simplify coordination and sharing of limited memory resources for complex integrated applications developed in the ECP/ATDM and ASC programs, and other multi-physics codes that combine many physics packages and libraries. The overarching goal of Umpire is to provide a single application-facing API for memory management on all platforms by decoupling memory operations from platform-specific memory spaces. It provides operations for: memory space discovery, memory provisioning, allocation and transfers, and introspection. Umpire development began in the Center for Applied Scientific Computing at Lawrence Livermore National Laboratory (LLNL). Umpire is funded primarily by LLNL ASC and ATMD portfolios and receives a small amount of support as part of ECP ST Project 2.3.1.06 STMP08-ISC. The source code is hosted and developed on GitHub. In this report, we explore possible integrations with other memory hierarchy management software. We examine both projects supported by ECP, as well as third-party and vendor provided tools.

97 MATHEMATICS AND COMPUTING↗

Evaluation of Accuracy and Performance of a Bilinear Interpolator with Dense EOS Tables in xRage

This study investigates the performance and accuracy of simulations using air and deuterium equation-of-state (EOS) tables that have been densified coupled with linear interpolation, as opposed to the default rational interpolation used in EOSPAC. EOS tables that are densified by adding in more temperature and density points that have been calculated by the rational function interpolator are able to improve the accuracy of the linear interpolator by over 50%. There was no change in runtime performance for a simulation using a dense EOS table with linear interpolation. In a memory limited case, dense EOS tables had no apparent effect on the simulation runtime or memory usage.

97 MATHEMATICS AND COMPUTING↗

Performance Models for Data Transfers: A Case Study with Molecular Chemistry Kernels

In distributed memory systems, it is paramount to develop strategies to overlap the data transfers between memory nodes with the computations in order to exploit the full potential of these systems. In this paper, we consider the problem of determining the order of data transfers between two memory nodes for a set of independent tasks with the objective of minimizing the makespan. We prove that, with limited memory capacity, the problem of obtaining the optimal order of data transfers is NP-complete. We propose several heuristics to determine this order and discuss the conditions that might be favorable to different heuristics. We analyze our heuristics on traces obtained by running 2 molecular chemistry kernels, namely, Hartree–Fock (HF) and Coupled Cluster Single Double (CCSD) on 10 nodes of an HPC system. Our results show that some of our heuristics achieve significant overlap for moderate memory capacities and resulting in makespans that are very close to the lower bound.

Kumar, Suraj↗

Streaming Matching and Edge Cover in Practice

Graph algorithms with polynomial space and time requirements often become infeasible for massive graphs with billions of edges or more. State-of-the-art approaches therefore employ approximate serial, parallel, and distributed algorithms to tackle these challenges. However, such approaches require storing the entire graph in memory and thus need access to costly computing resources such as clusters and supercomputers. In this paper, we present practical streaming approaches for solving massive graph problems using limited memory for two prototypical graph problems: maximum weighted matching and minimum weighted edge cover. For matching, we conduct a thorough computational study on two of the semi-streaming algorithms including a recent breakthrough result that achieves a $1/(2+\varepsilon)$-approximation of the weight while using $O( n \log W /\epsilon)$ memory (here $n$ is the number of vertices and $W$ is the maximum edge weight), designed by Paz and Schwartzman [SODA, 2017]. Empirically, we show that the semi-streaming algorithms produce matchings whose weight is close to the best $1/2$-approximate offline algorithm while requiring less time and an order-of-magnitude less memory. For minimum weighted edge cover, we develop three novel semi-streaming algorithms. Two of these algorithms require a single pass through the input graph, require $O(n \log n)$ memory, and provide a 2-approximation guarantee on the objective. We also leverage a relationship between approximate maximum weighted matching and approximate minimum weighted edge cover to develop a two-pass $3/2+\epsilon$-approximate algorithm with the memory requirement of Paz and Schwartzman's semi-streaming matching algorithm. These streaming approaches are compared against the state-of-the-art 3/2-approximate offline algorithm. The semi-streaming matching and the novel edge cover algorithms proposed in this paper can process graphs with several billions of edges in under 30 minutes using 6 GB of memory, which is at least an order of magnitude improvement from the offline (non-streaming) algorithms. For the largest graph, the best alternative offline parallel approximation algorithm (GPA+ROMA) could not finish in three hours even while employing hundreds of processors and 1 TB of memory. We also demonstrate an application of the semi-streaming algorithm by computing a matching using linearly bounded memory on item intersection graphs derived from three machine learning datasets, whereas the existing offline algorithms could not complete on one of these datasets since their memory requirements exceeded 1TB.

Ferdous, S M.↗

Distributed Tomographic Reconstruction with Quantization

Conventional tomographic reconstruction typically depends on centralized servers for both data storage and computation, leading to concerns about memory limitations and data privacy. Distributed reconstruction algorithms mitigate these issues by partitioning data across multiple nodes, reducing server load and enhancing privacy. However, these algorithms often encounter challenges related to memory constraints and communication overhead between nodes. In this paper, we introduce a decentralized Alternating Directions Method of Multipliers (ADMM) with configurable quantization. By distributing local objectives across nodes, our approach is highly scalable and can efficiently reconstruct images while adapting to available resources. To overcome communication bottlenecks, we propose two quantization techniques based on K-means clustering and JPEG compression. Numerical experiments with benchmark images illustrate the tradeoffs between communication efficiency, memory use, and reconstruction accuracy.

Miao, Runxuan↗

A memory-distributed quasi-Newton solver for nonlinear programming problems with a small number of general constraints

We approach the problem of parallelizing state-of-the-art nonlinear programming optimization algorithms.Specifically, we focus on parallelizing quasi-Newton interior-point methods that use limited-memory secant Hessian approximations. Such interior-point methods are known to have better convergence properties and to be more effective on large-scale problems than gradient-based and derivative-free optimization algorithms. We target nonlinear and potentially nonconvex optimization problems with an arbitrary number of bound constraints and a small number of general equality and inequality constraints on the optimization variables. These problems occur for example in the form of optimal control, optimal design, and inverse problems governed by ordinary or partial differential equations, whenever they are expressed in a “reduced-space” optimization approach. We introduce and analyze the time and space complexity of a decomposition method for solving the quasi-Newton linear systems that leverages the fact that the quasi-Newton Hessian matrix has a small number of dense blocks that border a low-rank update of a diagonal matrix. This enables an efficient parallelization on memory-distributed computers of the iterations of the optimization algorithm, a state-of-the-art filter line-search interior-point algorithm by Wächter et. al. We illustrate the efficiency of the proposed method by solving structural topology optimization problems on up to 4608 cores on a parallel machine.

97 MATHEMATICS AND COMPUTING↗