Search NASA⌕ Search

SEARCH · Search NASA

Results for “memory efficiency”

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 379 records · Page 21

Pseudo-random number generation for Brownian Dynamics and Dissipative Particle Dynamics simulations on GPU devices

Highlights: {yields} Molecular Dynamics codes implemented on GPUs have achieved two-order of magnitude computational accelerations. {yields} Brownian Dynamics and Dissipative Particle Dynamics simulations require a large number of random numbers per time step. {yields} We introduce a method for generating small batches of pseudorandom numbers distributed over many threads of calculations. {yields} With this method, Dissipative Particle Dynamics is implemented on a GPU device without requiring thread-to-thread communication. - Abstract: Brownian Dynamics (BD), also known as Langevin Dynamics, and Dissipative Particle Dynamics (DPD) are implicit solvent methods commonly used in models of soft matter and biomolecular systems. The interaction of the numerous solvent particles with larger particles is coarse-grained as a Langevin thermostat is applied to individual particles or to particle pairs. The Langevin thermostat requires a pseudo-random number generator (PRNG) to generate the stochastic force applied to each particle or pair of neighboring particles during each time step in the integration of Newton's equations of motion. In a Single-Instruction-Multiple-Thread (SIMT) GPU parallel computing environment, small batches of random numbers must be generated over thousands of threads and millions of kernel calls. In this communication we introduce a one-PRNG-per-kernel-call-per-thread scheme, in which a micro-stream of pseudorandom numbers is generated in each thread and kernel call. These high quality, statistically robust micro-streams require no global memory for state storage, are more computationally efficient than other PRNG schemes in memory-bound kernels, and uniquely enable the DPD simulation method without requiring communication between threads.

97 MATHEMATICS AND COMPUTING↗

Three-dimensional aerodynamic shape optimization using discrete sensitivity analysis

An aerodynamic shape optimization procedure based on discrete sensitivity analysis is extended to treat three-dimensional geometries. The function of sensitivity analysis is to directly couple computational fluid dynamics (CFD) with numerical optimization techniques, which facilitates the construction of efficient direct-design methods. The development of a practical three-dimensional design procedures entails many challenges, such as: (1) the demand for significant efficiency improvements over current design methods; (2) a general and flexible three-dimensional surface representation; and (3) the efficient solution of very large systems of linear algebraic equations. It is demonstrated that each of these challenges is overcome by: (1) employing fully implicit (Newton) methods for the CFD analyses; (2) adopting a Bezier-Bernstein polynomial parameterization of two- and three-dimensional surfaces; and (3) using preconditioned conjugate gradient-like linear system solvers. Whereas each of these extensions independently yields an improvement in computational efficiency, the combined effect of implementing all the extensions simultaneously results in a significant factor of 50 decrease in computational time and a factor of eight reduction in memory over the most efficient design strategies in current use. The new aerodynamic shape optimization procedure is demonstrated in the design of both two- and three-dimensional inviscid aerodynamic problems including a two-dimensional supersonic internal/external nozzle, two-dimensional transonic airfoils (resulting in supercritical shapes), three-dimensional transport wings, and three-dimensional supersonic delta wings. Each design application results in realistic and useful optimized shapes.

Burgreen, Gregory W.↗

Adaptive Multiresolution or Adaptive Mesh Refinement? A Case Study for 2D Euler Equations

We present adaptive multiresolution (MR) computations of the two-dimensional compressible Euler equations for a classical Riemann problem. The results are then compared with respect to accuracy and computational efficiency, in terms of CPU time and memory requirements, with the corresponding finite volume scheme on a regular grid. For the same test-case, we also perform computations using adaptive mesh refinement (AMR) imposing similar accuracy requirements. The results thus obtained are compared in terms of computational overhead and compression of the computational grid, using in addition either local or global time stepping strategies. We preliminarily conclude that the multiresolution techniques yield improved memory compression and gain in CPU time with respect to the adaptive mesh refinement method.

97 MATHEMATICS AND COMPUTING↗

Makalu: fast recoverable allocation of non-volatile memory

Byte addressable non-volatile memory (NVRAM) is likely to supplement, and perhaps eventually replace, DRAM. Applications can then persist data structures directly in memory instead of serializing them and storing them onto a durable block device. However, failures during execution can leave data structures in NVRAM unreachable or corrupt. In this paper, we present Makalu, a system that addresses non-volatile memory management. Makalu offers an integrated allocator and recovery-time garbage collector that maintains internal consistency, avoids NVRAM memory leaks, and is efficient, all in the face of failures. We show that a careful allocator design can support a less restrictive and a much more familiar programming model than existing persistent memory allocators. Our allocator significantly reduces the per allocation persistence overhead by lazily persisting non-essential metadata and by employing a post-failure recovery-time garbage collector. Experimental results show that the resulting online speed and scalability of our allocator are comparable to well-known transient allocators, and significantly better than state-of-the-art persistent allocators.

97 MATHEMATICS AND COMPUTING↗

Memory-optimized shift operator alternating direction implicit finite difference time domain method for plasma

Through introducing the alternating direction implicit (ADI) technique and the memory-optimized algorithm to the shift operator (SO) finite difference time domain (FDTD) method, the memory-optimized SO-ADI FDTD for nonmagnetized collisional plasma is proposed and the corresponding formulae of the proposed method for programming are deduced. In order to further the computational efficiency, the iteration method rather than Gauss elimination method is employed to solve the equation set in the derivation of the formulae. Complicated transformations and convolutions are avoided in the proposed method compared with the Z transforms (ZT) ADI FDTD method and the piecewise linear JE recursive convolution (PLJERC) ADI FDTD method. The numerical dispersion of the SO-ADI FDTD method with different plasma frequencies and electron collision frequencies is analyzed and the appropriate ratio of grid size to the minimum wavelength is given. The accuracy of the proposed method is validated by the reflection coefficient test on a nonmagnetized collisional plasma sheet. The testing results show that the proposed method is advantageous for improving computational efficiency and saving computer memory. The reflection coefficient of a perfect electric conductor (PEC) sheet covered by multilayer plasma and the RCS of the objects coated by plasma are calculated by the proposed method and the simulation results are analyzed.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Optimizing High Performance Markov Clustering for Pre-Exascale Architectures

HipMCL is a high-performance distributed memory implementation of the popular Markov Cluster Algorithm (MCL) and can cluster large-scale networks within hours using a few thousand CPU-equipped nodes. It relies on sparse matrix computations and heavily makes use of the sparse matrix-sparse matrix multiplication kernel (SpGEMM). The existing parallel algorithms in HipMCL are not scalable to Exascale architectures, both due to their communication costs dominating the runtime at large concurrencies and also due to their inability to take advantage of accelerators that are increasingly popular. In this work, we systematically remove scalability and performance bottlenecks of HipMCL. We enable GPUs by performing the expensive expansion phase of the MCL algorithm on GPU. Additionally, we propose a CPU-GPU joint distributed SpGEMM algorithm called pipelined Sparse SUMMA and integrate a probabilistic memory requirement estimator that is fast and accurate. Furthermore, we develop a new merging algorithm for the incremental processing of partial results produced by the GPUs, which improves the overlap efficiency and the peak memory usage. We also integrate a recent and faster algorithm for performing SpGEMM on CPUs. We validate our new algorithms and optimizations with extensive evaluations. With the enabling of the GPUs and integration of new algorithms, HipMCL is up to 12.4x faster, being able to cluster a network with 70 million proteins and 68 billion connections just under 15 minutes using 1024 nodes of ORNL's Summit supercomputer.

97 MATHEMATICS AND COMPUTING↗

Auxetic Thermo-Responsive Shape-Memory Polyurethane Aerogels

Shape memory polymer aerogels are stimuli-responsive aerogel materials that incorporate polymer aerogel characteristics such as flexibility at low density, and high porosity with shape-changing capabilities upon exposure to external stimuli. By designing the structure of these materials at meso- to macroscale, their shape memory response can be efficiently engineered to show an auxetic response such as a negative Poisson’s ratio, leading to an active and tunable metamaterial. We present a material technology based on a shape memory polyurethane aerogel derived from aliphatic triisocyanate and triethylene glycol, which demonstrates and remembers an auxetic shape change through a thermomechanical cycle first below, and then back again above the glass transition temperature of the polyurethane. Our auxetic shape memory polyurethane aerogels exhibit a significant auxetic response with a negative Poisson’s ratio of approximately -0.8 at 15% compressive strain. This technology is expected to impact multiple areas of interest in relation to commercial, aeronautic, and aerospace industries. Potential applications include minimally invasive medical technologies, soft robotics, and self-deployable structures and habitats. Structure-property relationships of the auxetic shape memory polymer aerogels will be discussed.

Sadeq Malakooti↗

A comparative study of two numerical approaches for solving Kim–Kim–Suzuki phase-field models

Among the standard multi-phase multi-component phase-field (PF) methods, the Kim–Kim–Suzuki (KKS) method has the advantage of decoupling interfacial energy from bulk energy and solving concentration as the conserved variable. There are two approaches to numerically solving a KKS method: the global solution approach (GSA) solves all variables in a global system simultaneously, and the local solution approach (LSA) solves phase concentrations locally using a Newton solver. This work compares the performance of LSA and GSA for solving four KKS models of increasing complexity with the finite element method using the MOOSE framework. The solution accuracy, degrees of freedom (DOFs), memory usage, and computational efficiency are compared. We find that GSA and LSA generate similar solutions, with a maximum difference of only 0.34%. For each model, LSA has a lower number of DOFs, utilizes less memory, and less wall time. Additionally, the savings of memory and wall time in LSA increase with increasing mesh density of the same model and are more pronounced in models with higher dimensionality and more nodes. However, GSA is easier to implement in existing codes and can better solve highly nonlinear systems by utilizing sophisticated solvers.

36 MATERIALS SCIENCE↗

Distributed non-equilibrium Green’s function algorithms for the simulation of nanoelectronic devices with scattering

Through the non-equilibrium Green’s function (NEGF) formalism, quantum-scale device simulation can be performed with the inclusion of electron-phonon scattering. Yet, the simulation of realistically sized devices under the NEGF formalism typically requires prohibitive amounts of memory and computation time. Two of the most demanding computational problems for NEGF simulation involve mathematical operations with structured matrices called semiseparable matrices. In this work, we present parallel approaches for these computational problems which allow for efficient distribution of both memory and computation based upon the underlying device structure. This is vital when simulating realistically sized devices due to the aforementioned computational burdens. First, we consider determining a distributed compact representation for the retarded Green’s function matrix G R . This compact representation is exact and allows for any entry in the matrix to be generated through the inherent semiseparable structure. The second parallel operation allows for the computation of electron density and current characteristics for the device. Specifically, matrix products between the distributed representation for the semiseparable matrix G R and the self-energy scattering terms in Σ < produce the less-than Green’s function G < . As an illustration of the computational efficiency of our approach, we stably generate the mobility for nanowires with cross-sectional sizes as large as 4.5 nm, assuming an atomistic model with scattering.

47 OTHER INSTRUMENTATION↗

Efficient checkpointing schemes for depletion perturbation solutions on memory-limited architectures

We describe a methodology for decreasing the memory footprint and machine I/O load associated with the need to access a forward solution during an adjoint solve. Specifically, we are interested in the depletion perturbation equations, where terms in the adjoint Bateman and transport equations depend on the forward flux solution. Checkpointing is the procedure of storing snapshots of the forward solution to disk and using these snapshots to recompute the parts of the forward solution that are necessary for the adjoint solve. For large problems, however, the storage cost of just a few copies of an angular flux vector can exceed the available RAM on the host machine. We propose a methodology that does not checkpoint the angular flux vector; instead, we write and store converged source moments, which are typically of a much lower dimension than the angular flux solution. This reduces the memory footprint and I/O load of the problem, but requires that we perform single sweeps to reconstruct flux vectors on demand. We argue that this trade-off is exactly the kind of algorithm that will scale on advanced, memory-limited architectures. We analyze the cost, in terms of FLOPS and memory footprint, of five checkpointing schemes. We also provide computational results that support the analysis and show that the memory-for-work trade off does improve time to solution. (authors)

97 MATHEMATICS AND COMPUTING↗

Low-temperature processed beta-phase In 2 Se 3 ferroelectric semiconductor thin film transistors

As scaling becomes increasingly difficult, there is growing interest in vertical or three-dimensional stacking of transistors and especially memory. Ferroelectric semiconductor field effect transistors can be key enablers to improve energy efficiency and overall chip and memory performance. In this work, low-temperature processed, back-end-of-the-line compatible transistors were demonstrated by depositing a layered chalcogenide ferroelectric semiconductor, beta-phase In 2 Se 3 , at temperature as low as 400 °C. Top gate n-channel In 2 Se 3 thin film transistors were fabricated with field-effect mobility ∼1 cm 2 V −1 s −1 , and simple polarization switching based memory results are presented.

36 MATERIALS SCIENCE↗

GentenMPI: Distributed Memory Sparse Tensor Decomposition

GentenMPl is a toolkit of sparse canonical polyadic (CP) tensor decomposition algorithms that is designed to run effectively on distributed-memory high-performance computers. Its use of distributed-memory parallelism enables it to efficiently decompose tensors that are too large for a single compute node's memory. GentenMPl leverages Sandia's decades-long investment in the Trilinos solver framework for much of its parallel-computation capability. Trilinos contains numerical algorithms and linear algebra classes that have been optimized for parallel simulation of complex physical phenomena. This work applies these tools to the data science problem of sparse tensor decomposition. In this report, we describe the use of Trilinos in GentenMPl, extensions needed for sparse tensor decomposition, and implementations of the CP-ALS (CP via alternating least squares) and GCP-SGD (generalized CP via stochastic gradient descent) sparse tensor decomposition algorithms. We show that GentenMPl can decompose sparse tensors of extreme size, e.g., a 12.6-terabyte tensor on 8192 computer cores. We demonstrate that the Trilinos backbone provides good strong and weak scaling of the tensor decomposition algorithms.

97 MATHEMATICS AND COMPUTING↗

Distributed-Memory Sparse Deep Neural Network Inference Using Global Arrays

Partitioned Global Address Space (PGAS) models exhibit tremendous promise in developing efficient and productive distributed-memory parallel applications. They have been used extensively in scientific computations due to conveniently offering a ``shared-memory''-like model and convenient interfaces that separate communication with synchronization. Traditionally, PGAS communication models have been applied to dense/contiguously distributed data, but most modern applications depict varied levels of sparsity. Existing PGAS models require certain adaptations to support distributed sparse computations, since associated computations often require matrix arithmetic, in addition to data movement. The Global Arrays toolkit from Pacific Northwest National Laboratory (PNNL) is one of the earliest PGAS models to combine one-sided data communication and distributed matrix operations and is still used in the popular NWChem quantum chemistry suite. Recently, we have expanded the Global Arrays toolkit to support common sparse operations, like sparse matrix-dense matrix multiplies (SpMM), sparse matrix-sparse matrix multiplication (SpGEMM) and Sampled Dense-Dense Matrix Multiplication (SDDMM). As it turns out, these operations are the bedrock of sparse Deep Learning (DL); sparse deep neural networks and Graph Neural Networks (GNNs) have gained increasing attention recently in achieving speedups on training and inference with reduced memory footprints. Unlike scientific applications in High Performance Computing (HPC), modern (distributed-memory capable) DL toolkits often rely on non-standardized and closed-source vendor software optimizations, creating challenges in software-hardware co-design at scale. Our goal is to support a variety of distributed-memory sparse matrix operations and helper functions in the newly created Sparse Global Arrays (SGA), such that it is possible to build portable and productive Machine Learning scenarios for algorithm/software and hardware codesign purposes. Contemporary data-parallel schemes for training/inference are undergoing a major overhaul since model replication limits scalability and causes resource inefficiencies. As such, we have adopted tensor parallelism in decomposing the model and inputs, to mitigate memory issues. Current implementation is built on top of MPI and uses CPUs to maximize the portability across the platforms.

Distributed computing, machine learning↗

Advanced compilation techniques in the PARADIGM compiler for distributed-memory multicomputers

The PARADIGM compiler project provides an automated means to parallelize programs, written in a serial programming model, for efficient execution on distributed-memory multicomputers. .A previous implementation of the compiler based on the PTD representation allowed symbolic array sizes, affine loop bounds and array subscripts, and variable number of processors, provided that arrays were single or multi-dimensionally block distributed. The techniques presented here extend the compiler to also accept multidimensional cyclic and block-cyclic distributions within a uniform symbolic framework. These extensions demand more sophisticated symbolic manipulation capabilities. A novel aspect of our approach is to meet this demand by interfacing PARADIGM with a powerful off-the-shelf symbolic package, Mathematica. This paper describes some of the Mathematica routines that performs various transformations, shows how they are invoked and used by the compiler to overcome the new challenges, and presents experimental results for code involving cyclic and block-cyclic arrays as evidence of the feasibility of the approach.

Su, Ernesto↗

Feedback-directed page placement for ccNUMA via hardware-generated memory traces

Non-uniform memory architectures with cache coherence (ccNUMA) are becoming increasingly common, not just for large-scale high performance platforms but also in the context of multi-core architectures. Under ccNUMA, data placement may influence overall application performance significantly as references resolved locally to a processor/core impose lower latencies than remote ones. This work develops a novel hardware-assisted page placement paradigm based on automated tracing of the memory references made by application threads. Two placement schemes, modeling both singlelevel and multi-level latencies, allocate pages near processors that most frequently access that memory page. These schemes leverage performance monitoring capabilities of contemporary microprocessors to efficiently extract an approximate trace of memory accesses. This information is used to decide page affinity, i.e., the node to which the page is bound. The method operates entirely in user space, is widely automated, and handles not only static but also dynamic memory allocation. Experiments show that this method, although based on lossy tracing, can efficiently and effectively improve page placement, leading to an average wall-clock execution time saving of over 20% for the tested benchmarks on the SGI Altix with a 2x remote access penalty and 12% on AMD Opterons with a 1.3–2.0x access penalty. This is accompanied by a one-time tracing overhead of 2.7% over the overall original program wallclock time.

Computer Science↗

Progress toward picosecond on-chip magnetic memory

Here, we offer a perspective on the prospects of ultrafast spintronics and opto-magnetism as a pathway to high-performance, energy-efficient, and non-volatile embedded memory in digital integrated circuit applications. Conventional spintronic devices, such as spin-transfer-torque magnetic-resistive random-access memory (STT-MRAM) and spin–orbit torque MRAM, are promising due to their non-volatility, energy-efficiency, and high endurance. STT-MRAMs are now entering into the commercial market; however, they are limited in write speed to the nanosecond timescale. Improvement in the write speed of spintronic devices can significantly increase their usefulness as viable alternatives to the existing CMOS-based devices. In this article, we discuss recent studies that advance the field of ultrafast spintronics and opto-magnetism. An optimized ferromagnet–ferrimagnet exchange-coupled magnetic stack, which can serve as the free layer of a magnetic tunnel junction (MTJ), can be optically switched in as fast as ~3 ps. Integration of ultrafast magnetic switching of a similar stack into an MTJ device has enabled electrical readout of the switched state using a relatively larger tunneling magnetoresistance ratio. Purely electronic ultrafast spin–orbit torque induced switching of a ferromagnet has been demonstrated using ~6 ps long charge current pulses. We conclude our Perspective by discussing some of the challenges that remain to be addressed to accelerate ultrafast spintronics technologies toward practical implementation in high-performance digital information processing systems.

42 ENGINEERING↗

Large-scale inverse model analyses employing fast randomized data reduction: INVERSE MODELING WITH DATA REDUCTION

When the number of observations is large, it is computationally challenging to apply classical inverse modeling techniques. We have developed a new computationally efficient technique for solving inverse problems with a large number of observations (e.g., on the order of 10 7 or greater). Our method, which we call the randomized geostatistical approach (RGA), is built upon the principal component geostatistical approach (PCGA). We employ a data reduction technique combined with the PCGA to improve the computational efficiency and reduce the memory usage. Specifically, we employ a randomized numerical linear algebra technique based on a so-called “sketching” matrix to effectively reduce the dimension of the observations without losing the information content needed for the inverse analysis. In this way, the computational and memory costs for RGA scale with the information content rather than the size of the calibration data. Our algorithm is coded in Julia and implemented in the MADS open-source high-performance computational framework (http://mads.lanl.gov). We apply our new inverse modeling method to invert for a synthetic transmissivity field. Compared to a standard geostatistical approach (GA), our method is more efficient when the number of observations is large. Moreover, our method is capable of solving larger inverse problems than the standard GA and PCGA approaches. Thus, our new model inversion method is a powerful tool for solving large-scale inverse problems. The method can be applied in any field and is not limited to hydrogeological applications such as the characterization of aquifer heterogeneity.

54 ENVIRONMENTAL SCIENCES↗

Initial CEED API (CEED-MS10)

Key components of CEED software development involve fast finite element operator storage and evaluation, architecture optimizations, performant algorithms for all orders, global kernels for finite element operators, efficient use of the memory sub-system, optimal data locality and motion, enhanced scalability and parallelism, and fast tensor contractions. As ultimate goals of the project, CEED is exploring and identifying the best algorithms for the full range of discretizations and applying algorithmic and software development to support ECP applications’ needs. In this milestone, we initiated the development and implementation of a common CEED API consisting of two main components: low-level and high-level APIs. We accomplished the goal of having the low-level component ready for the release of CEED v1.0 in the next quarter, see CEED-MS13. In this report we also provide a highlight from CEED’s integration with the MARBL app and the corresponding Laghos miniapp. A main theme in this work has been the systematic transition of advanced algorithms, developed and tested in Laghos, to the large-scale multi-physics settings of the MARBL code. The low-level CEED API provides a set of FE kernels and components for writing new low-level kernels. Examples include: vector and sparse linear algebra, element matrix assembly over a batch of elements, partial assembly and action for efficient high-order operators like mass, diffusion, advection, etc. The main goal of the low-level API is to establish the basis for the high-level API. Also, identifying such low-level kernels and providing a reference implementation for them serves as the basis for collaboration with vendors and STs. Another goal for the low-level API is to make it easy for applications with their own discretization infrastructure, who do not want to switch directly to using the high-level CEED API, to still evaluate and use the core operations provided by the low-level API. One of the challenges with high-order methods is that a global sparse matrix is no longer a good representation of a high-order linear operator, both with respect to the FLOPs needed for its evaluation, as well as the memory transfer needed for a matvec. Thus, another important question we tried to address while developing the low-level API is: is it possible to abstract out in a natural and efficient way the different implementations and data structures that arise when dealing with various computational device types: CPUs, GPUs, etc. Fully answering this question is beyond the scope of this particular milestone, since that requires longer time as we develop and expand the devices we support. The future high-level API will be the frontend provided by CEED consisting of high-level abstractions such as meshes, finite element spaces and solutions, bilinear forms, solvers, etc. At this high-level, an important aspect of the API will be to provide a uniform abstraction layer for writing code in a portable manner. To achieve that, we plan to utilize the MPI programming model combined with a device abstraction layer (DAL) for finite element computations that we will develop. The DAL will serve as the bridge between the low- and the high-level APIs, allowing a variety of programming models to be used for the low-level kernel implementations. The software artifacts delivered as part of this milestone include an initial CEED API software implementation (libCEED) provided through the CEED website, http://ceed.exascaleproject.org and the CEED GitHub organization, http://github.com/ceed. In addition to details and results from the above R&D efforts, in this document we are also reporting on other project-wide activities performed in Q1 of FY18, including: the Nek5000 v17.0 MFEM v3.3.2 and PETSc v3.8 releases, collaborations with ECP/ST and SciDAC projects, the inaugural Nek5000 hackathon and various outreach activities.

97 MATHEMATICS AND COMPUTING↗