Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “compute in 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 1,639 records · Page 91

Annular Flow Simulation Supported by Iterative In-Memory Mesh Adaptation

Various flow regimes exist in a boiling water reactor (BWR) as the steam quality increases in the uprising coolant flow, from bubbly flow, slug/churn flow, to annular flow. The annular flow is characterized by the presence of a fast-moving gas core and the surrounding liquid film flowing on the conduit wall. Additionally, entrained droplets can be observed in the gas core with ingested bubbles in the liquid film. The dynamics occurring on the wavy interface between the liquid film and gas core plays a crucial role in affecting the heat transfer rate and pressure drop within the BWR core. However, a fundamental understanding of annular flow is still lacking, partly due to the difficulty in obtaining detailed local data in annular flow experiments. In the current study, a novel simulation framework is developed for the annular flow by coupling a computational fluid dynamics flow solver with state-of-the-art meshing software. The gas-liquid interface is tracked with the level set method. Based on the computed flow solutions, the computational mesh is dynamically adapted in memory to meet the local mesh resolution requirement. This iterative simulation-adaptation framework can ensure the fine mesh resolution across the interface, which not only helps mitigate the mass conservation degradation known to level set methods but also improves the representation of dramatic interface topological changes such as wave breaking and droplet entrainment. The present investigation will shed light onto the complex interfacial processes involved in annular flow and generate much needed simulation data for annular flow modeling.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

A case for Virtual Machine based Fault Injection in a High-Performance Computing Environment

Large-scale computing platforms provide tremendous capabilities for scientific discovery. These systems have hundreds of thousands of computing cores, hundreds of terabytes of memory, and enormous high-performance interconnection networks. These systems are facing enormous challenges to achieve performance at such scale. Failures are an Achilles heel of these enormous systems. As applications and system software scale up to multi-petaflop and beyond to exascale platforms, the occurrence of failure will be much more common. This has given rise to a push in fault-tolerance and resilience research for HPC systems. This includes work on log analysis to identify types of failures, enhancements to the Message Passing Interface (MPI) to incorporate fault awareness, and a variety of fault tolerance mechanisms that span redundant computation, algorithm based fault tolerance, and advanced checkpoint/restart techniques. While there is much work to be done on the FT/Resilience mechanisms for such large-scale systems, there is also a profound gap in the tools for experimentation. This gap is compounded by the fact that HPC environments have stringent performance requirements and are often highly customized. The tool chain for these systems are often tailored for the platform and while the majority of systems on the Top500 Supercomputer list run Linux, these operating environments typically contain many site/machine specific enhancements. Therefore, it is desirable to maintain a consistent execution environment to minimize end-user (scientist) interruption. The work on system-level virtualization for HPC system offers a unique opportunity to maintain a consistent execution environment via a virtual machine (VM). Recent work on virtualization for HPC has shown that low-overhead, high performance systems can be realized. Virtualization also provides a clean abstraction for building experimental tools for investigation into the effects of failures in HPC and the related research on FT/Resilience mechanisms and policies. In this paper we discuss the motivation for tools to perform fault injection in an HPC context, and outline an approach that can leverage virtualization.

99 GENERAL AND MISCELLANEOUS↗

Unified Memory: GPGPU-Sim/UVM Smart Integration

CPU/GPU heterogeneous compute platforms are an ubiquitous element in computing and a programming model specified for this heterogeneous computing model is important for both performance and programmability. A programming model that exposes the shared, unified, address space between the heterogeneous units is a necessary step in this direction as it removes the burden of explicit data movement from the programmer while maintaining performance. GPU vendors, such as AMD and NVIDIA, have released software-managed runtimes that can provide programmers the illusion of unified CPU and GPU memory by automatically migrating data in and out of the GPU memory. However, this runtime support is not included in GPGPU-Sim, a commonly used framework that models the features of a modern graphics processor that are relevant to non-graphics applications. UVM Smart was developed, which extended GPGPU-Sim 3.x to in- corporate the modeling of on-demand pageing and data migration through the runtime. This report discusses the integration of UVM Smart and GPGPU-Sim 4.0 and the modifications to improve simulation performance and accuracy.

97 MATHEMATICS AND COMPUTING↗

Portable Electromyograph

A portable electronic apparatus records electromyographic (EMG) signals in as many as 16 channels at a sampling rate of 1,024 Hz in each channel. The apparatus (see figure) includes 16 differential EMG electrodes (each electrode corresponding to one channel) with cables and attachment hardware, reference electrodes, an input/output-and-power-adapter unit, a 16-bit analog-to-digital converter, and a hand-held computer that contains a removable 256-MB flash memory card. When all 16 EMG electrodes are in use, full-bandwidth data can be recorded in each channel for as long as 8 hours. The apparatus is powered by a battery and is small enough that it can be carried in a waist pouch. The computer is equipped with a small screen that can be used to display the incoming signals on each channel. Amplitude and time adjustments of this display can be made easily by use of touch buttons on the screen. The user can also set up a data-acquisition schedule to conform to experimental protocols or to manage battery energy and memory efficiently. Once the EMG data have been recorded, the flash memory card is removed from the EMG apparatus and placed in a flash-memory- card-reading external drive unit connected to a personal computer (PC). The PC can then read the data recorded in the 16 channels. Preferably, before further analysis, the data should be stored in the hard drive of the PC. The data files are opened and viewed on the PC by use of special- purpose software. The software for operation of the apparatus resides in a random-access memory (RAM), with backup power supplied by a small internal lithium cell. A backup copy of this software resides on the flash memory card. In the event of loss of both main and backup battery power and consequent loss of this software, the backup copy can be used to restore the RAM copy after power has been restored. Accessories for this device are also available. These include goniometers, accelerometers, foot switches, and force gauges.

De Luca, Gianluca↗

Efficient derivative computation for unsteady fatigue-constrained nonlinear aero-structural wind turbine blade optimization

Gradient-based optimization offers significant efficiency advantages for wind turbine blade design, but its application has often been limited by the cost and accuracy of finite-difference derivative calculations, especially when fatigue constraints are considered. In this work, we systematically compare and evaluate four differentiation techniques, namely algorithmic differentiation, implicit differentiation, sparsity exploitation, and parallelization, to determine their effectiveness in computing accurate gradients through time-domain aero-structural simulations. By integrating these techniques with unsteady nonlinear aerodynamic and structural models, we develop software designed for accurate gradient computation. We show that combining these techniques addresses memory and runtime challenges associated with long simulations required by design load cases. Specifically, the most effective combination reduces derivative computation wall time by over an order of magnitude compared to finite differencing while maintaining superior accuracy. We demonstrate this approach in a proof-of-concept aero-structural optimization of a wind turbine blade that improves the cost of energy by 12.78 %. This comparative study establishes a viable approach for fatigue-aware blade design that balances computational efficiency with modeling accuracy.

17 WIND ENERGY↗

Method and system for data clustering for very large databases

Multi-dimensional data contained in very large databases is efficiently and accurately clustered to determine patterns therein and extract useful information from such patterns. Conventional computer processors may be used which have limited memory capacity and conventional operating speed, allowing massive data sets to be processed in a reasonable time and with reasonable computer resources. The clustering process is organized using a clustering feature tree structure wherein each clustering feature comprises the number of data points in the cluster, the linear sum of the data points in the cluster, and the square sum of the data points in the cluster. A dense region of data points is treated collectively as a single cluster, and points in sparsely occupied regions can be treated as outliers and removed from the clustering feature tree. The clustering can be carried out continuously with new data points being received and processed, and with the clustering feature tree being restructured as necessary to accommodate the information from the newly received data points.

Zhang, Tian↗

SharP: Towards Programming Extreme-Scale Systems with Hierarchical Heterogeneous Memory

The pre-exascale systems are expected to have a significant amount of hierarchical and heterogeneous on-node memory, and this trend of system architecture in extreme-scale systems is expected to continue into the exascale era. Along with hierarchical-heterogeneous memory, the system typically has a high-performing network and a compute accelerator. This system architecture is not only effective for running traditional High Performance Computing (HPC) applications (Big-Compute), but also running data-intensive HPC applications and Big-Data applications. As a consequence, there is a growing desire to have a single system serve the needs of both Big-Compute and Big-Data applications. Though the system architecture supports the convergence of the Big-Compute and Big-Data, the programming models have yet to evolve to support either hierarchical-heterogeneous memory systems or the convergence. In this work, we propose and develop the programming abstraction called SHARed data-structure centric Programming abstraction (SharP) to address both of these goals, i.e., provide (1) a simple, usable, and portable abstraction for hierarchical-heterogeneous memory and (2) a unified programming abstraction for Big-Compute and Big-Data applications. To evaluate SharP, we implement a Stencil benchmark using SharP, port QMCPack, a petascale-capable application, and adapt Memcached ecosystem, a popular Big-Data framework, to use SharP, and quantify the performance and productivity advantages. Additionally, we demonstrate the simplicity of using SharP on different memories including DRAM, High-bandwidth Memory (HBM), and non-volatile random access memory (NVRAM).

97 MATHEMATICS AND COMPUTING↗

Roofline model toolkit: A practical tool for architectural and program analysis

We present preliminary results of the Roofline Toolkit for multicore, many core, and accelerated architectures. This paper focuses on the processor architecture characterization engine, a collection of portable instrumented micro benchmarks implemented with Message Passing Interface (MPI), and OpenMP used to express thread-level parallelism. These benchmarks are specialized to quantify the behavior of different architectural features. Compared to previous work on performance characterization, these microbenchmarks focus on capturing the performance of each level of the memory hierarchy, along with thread-level parallelism, instruction-level parallelism and explicit SIMD parallelism, measured in the context of the compilers and run-time environments. We also measure sustained PCIe throughput with four GPU memory managed mechanisms. By combining results from the architecture characterization with the Roofline model based solely on architectural specifications, this work offers insights for performance prediction of current and future architectures and their software systems. To that end, we instrument three applications and plot their resultant performance on the corresponding Roofline model when run on a Blue Gene/Q architecture.

97 MATHEMATICS AND COMPUTING↗

Implementing and Evaluating Multithreaded Triad Census Algorithms on the Cray XMT

Commonly represented as directed graphs, social networks depict relationships and behaviors among social entities such as people, groups, and organizations. Social network analysis denotes a class of mathematical and statistical methods designed to study and measure social networks. Beyond sociolo-gy, social network analysis methods are being applied to other types of data in other domains such as bioinformatics, computer networks, national security, and economics. For particular problems, the size of a social network can grow to millions of nodes and tens of millions of edges or more. In such cases, researchers could benefit from the application of social network analysis algorithms on high-performance architectures and systems. The Cray XMT is a third generation multithreaded system based on the Cray XT-3/4 platform. Like most other multithreaded architectures, the Cray XMT is designed to tolerate memory access latencies by switching context between threads. The processors maintain multiple threads of execution and util-ize hardware-based context switching to overlap the memory latency incurred by any thread with the computations from other threads. Due to its memory latency tolerance, the Cray XMT has the poten-tial of significantly improving the execution speed of irregular data-intensive applications such as those found in social network analysis. In this paper, we describe our experiences in developing and optimizing three implementations of a social network analysis method known as triadic analysis to execute on the Cray XMT. The three im-plementations possess different execution complexities, qualities, and characteristics. We evaluate how the various attributes of the codes affect their performance on the Cray XMT. We also explore the effects of different compiler options and execution strategies on the different triadic analysis im-plementations and identify general XMT programming issues and lessons learned.

45 MILITARY TECHNOLOGY, WEAPONRY, AND NATIONAL DEF↗

Architecture-Aware Models of AI Engines for High-Performance Matrix Matrix Multiplication

The AI Engine (AIE) architecture, available in systems from mobile SoCs to server-class FPGAs, aims to efficiently execute AI/ML tasks through a two-dimensional array of compute tiles. Previous work on AIEs has explored different approaches to mapping computation across spatial arrays, but the compute kernel running on each tile has not been the focus. Additionally, the AIE-ML architecture introduces memory tiles and omits programmable logic, requiring new approaches to staging and moving data throughout the array. In this work we update analytical models developed for CPUs to produce the design of high performance kernels while introducing new model considerations such as memory structure, throughput, and latency as required by the AIE hardware. We evaluate our models by developing AIE-ML kernels for matrix multiplication in low-precision data types showing performance up to 95% of compute peak for the kernel when data resides in local memory and above 90% of compute peak when data resides in main memory.

Binder, Elliott D. [Carnegie Mellon University, Pi↗

Optimal renormalization of multiscale systems

While model order reduction is a promising approach in dealing with multi-scale time-dependent systems that are too large or too expensive to simulate for long times, the resulting reduced order models can suffer from instabilities. We have recently developed a time-dependent renormalization approach to stabilize such reduced models. In the current work, we extend this framework by introducing a parameter that controls the time-decay of the memory of such models and optimally selecting this parameter based on limited fully resolved simulations. First, we demonstrate our framework on the inviscid Burgers equation whose solution develops a finite-time singularity. Our renormalized reduced order models are stable and accurate for long times while using for their calibration only data from a full order simulation before the occurrence of the singularity. Furthermore, we apply this framework to the 3D Euler equations of incompressible fluid flow, where the problem of finite-time singularity formation is still open and where brute force simulation is only feasible for short times. Our approach allows us to obtain for the first time a perturbatively renormalizable model which is stable for long times and includes all the complex effects present in the 3D Euler dynamics. We find that, in each application, the renormalization coefficients display algebraic decay with increasing resolution, and that the parameter which controls the time-decay of the memory is problem-dependent.

Price, Jacob↗

Decomposition Algorithm for Global Reachability on a Time-Varying Graph

A decomposition algorithm has been developed for global reachability analysis on a space-time grid. By exploiting the upper block-triangular structure, the planning problem is decomposed into smaller subproblems, which is much more scalable than the original approach. Recent studies have proposed the use of a hot-air (Montgolfier) balloon for possible exploration of Titan and Venus because these bodies have thick haze or cloud layers that limit the science return from an orbiter, and the atmospheres would provide enough buoyancy for balloons. One of the important questions that needs to be addressed is what surface locations the balloon can reach from an initial location, and how long it would take. This is referred to as the global reachability problem, where the paths from starting locations to all possible target locations must be computed. The balloon could be driven with its own actuation, but its actuation capability is fairly limited. It would be more efficient to take advantage of the wind field and ride the wind that is much stronger than what the actuator could produce. It is possible to pose the path planning problem as a graph search problem on a directed graph by discretizing the spacetime world and the vehicle actuation. The decomposition algorithm provides reachability analysis of a time-varying graph. Because the balloon only moves in the positive direction in time, the adjacency matrix of the graph can be represented with an upper block-triangular matrix, and this upper block-triangular structure can be exploited to decompose a large graph search problem. The new approach consumes a much smaller amount of memory, which also helps speed up the overall computation when the computing resource has a limited physical memory compared to the problem size.

Kuwata, Yoshiaki↗

Multi-GPU implementation of a VMAT treatment plan optimization algorithm

Purpose: Volumetric modulated arc therapy (VMAT) optimization is a computationally challenging problem due to its large data size, high degrees of freedom, and many hardware constraints. High-performance graphics processing units (GPUs) have been used to speed up the computations. However, GPU’s relatively small memory size cannot handle cases with a large dose-deposition coefficient (DDC) matrix in cases of, e.g., those with a large target size, multiple targets, multiple arcs, and/or small beamlet size. The main purpose of this paper is to report an implementation of a column-generation-based VMAT algorithm, previously developed in the authors’ group, on a multi-GPU platform to solve the memory limitation problem. While the column-generation-based VMAT algorithm has been previously developed, the GPU implementation details have not been reported. Hence, another purpose is to present detailed techniques employed for GPU implementation. The authors also would like to utilize this particular problem as an example problem to study the feasibility of using a multi-GPU platform to solve large-scale problems in medical physics. Methods: The column-generation approach generates VMAT apertures sequentially by solving a pricing problem (PP) and a master problem (MP) iteratively. In the authors’ method, the sparse DDC matrix is first stored on a CPU in coordinate list format (COO). On the GPU side, this matrix is split into four submatrices according to beam angles, which are stored on four GPUs in compressed sparse row format. Computation of beamlet price, the first step in PP, is accomplished using multi-GPUs. A fast inter-GPU data transfer scheme is accomplished using peer-to-peer access. The remaining steps of PP and MP problems are implemented on CPU or a single GPU due to their modest problem scale and computational loads. Barzilai and Borwein algorithm with a subspace step scheme is adopted here to solve the MP problem. A head and neck (H and N) cancer case is then used to validate the authors’ method. The authors also compare their multi-GPU implementation with three different single GPU implementation strategies, i.e., truncating DDC matrix (S1), repeatedly transferring DDC matrix between CPU and GPU (S2), and porting computations involving DDC matrix to CPU (S3), in terms of both plan quality and computational efficiency. Two more H and N patient cases and three prostate cases are used to demonstrate the advantages of the authors’ method. Results: The authors’ multi-GPU implementation can finish the optimization process within ∼1 min for the H and N patient case. S1 leads to an inferior plan quality although its total time was 10 s shorter than the multi-GPU implementation due to the reduced matrix size. S2 and S3 yield the same plan quality as the multi-GPU implementation but take ∼4 and ∼6 min, respectively. High computational efficiency was consistently achieved for the other five patient cases tested, with VMAT plans of clinically acceptable quality obtained within 23–46 s. Conversely, to obtain clinically comparable or acceptable plans for all six of these VMAT cases that the authors have tested in this paper, the optimization time needed in a commercial TPS system on CPU was found to be in an order of several minutes. Conclusions: The results demonstrate that the multi-GPU implementation of the authors’ column-generation-based VMAT optimization can handle the large-scale VMAT optimization problem efficiently without sacrificing plan quality. The authors’ study may serve as an example to shed some light on other large-scale medical physics problems that require multi-GPU techniques.

60 APPLIED LIFE SCIENCES↗

MetaStrider: Architectures for Scalable Memory-centric Reduction of Sparse Data Streams

Reduction is an operation performed on the values of two or more key-value pairs that share the same key. Reduction of sparse data streams finds application in a wide variety of domains such as data and graph analytics, cybersecurity, machine learning, and HPC applications. However, these applications exhibit low locality of reference, rendering traditional architectures and data representations inefficient. This article presents MetaStrider, a significant algorithmic and architectural enhancement to the state-of-the-art, SuperStrider. Furthermore, these enhancements enable a variety of parallel, memory-centric architectures that we propose, resulting in demonstrated performance that scales near-linearly with available memory-level parallelism.

97 MATHEMATICS AND COMPUTING↗

Basic Requirements for Systems Software Research and Development

Our success over the past ten years evaluating and developing advanced computing technologies has been due to a simple research and development (R/D) model. Our model has three phases: (a) evaluating the state-of-the-art, (b) identifying problems and creating innovations, and (c) developing solutions, improving the state- of-the-art. This cycle has four basic requirements: a large production testbed with real users, a diverse collection of state-of-the-art hardware, facilities for evalua- tion of emerging technologies and development of innovations, and control over system management on these testbeds. Future research will be irrelevant and future products will not work if any of these requirements is eliminated. In order to retain our effectiveness, the numerical aerospace simulator (NAS) must replace out-of-date production testbeds in as timely a fashion as possible, and cannot afford to ignore innovative designs such as new distributed shared memory machines, clustered commodity-based computers, and multi-threaded architectures.

Kuszmaul, Chris↗

Vibrational heat-bath configuration interaction with semistochastic perturbation theory using harmonic oscillator or VSCF modals

Vibrational heat-bath configuration interaction (VHCI)—a selected configuration interaction technique for vibrational structure theory—has recently been developed in two independent works [J. H. Fetherolf and T. C. Berkelbach, J. Chem. Phys. 154, 074104 (2021); A. U. Bhatty and K. R. Brorsen, Mol. Phys. 119, e1936250 (2021)], where it was shown to provide accuracy on par with the most accurate vibrational structure methods with a low computational cost. Here, we eliminate the memory bottleneck of the second-order perturbation theory correction using the same (semi)stochastic approach developed previously for electronic structure theory. This allows us to treat, in an unbiased manner, much larger perturbative spaces, which are necessary for high accuracy in large systems. Stochastic errors are easily controlled to be less than 1 cm−1. We also report two other developments: (i) we propose a new heat-bath criterion and an associated exact implicit sorting algorithm for potential energy surfaces expressible as a sum of products of one-dimensional potentials; (ii) we formulate VHCI to use a vibrational self-consistent field (VSCF) reference, as opposed to the harmonic oscillator reference configuration used in previous reports. Our tests are done with quartic and sextic force fields, for which we find that with VSCF, the minor improvements to accuracy are outweighed by the higher computational cost associated the matrix element evaluations. We expect VSCF-based VHCI to be important for more general potential representations, for which the harmonic oscillator basis function integrals are no longer analytic.

Chemistry↗

On beyond the standard model for high explosives: challenges & obstacles to surmount

Plastic-bonded explosives (PBX) are heterogeneous materials. Nevertheless, current explosive models treat them as homogeneous materials. To compensate, an empirically determined effective burn rate is used in place of a chemical reaction rate. A significant limitation of these models is that different burn parameters are needed for applications in different regimes; for example, shock initiation of a PBX at different initial temperatures or different initial densities. This is due to temperature fluctuations generated when a heterogeneous material is shock compressed. Localized regions of high temperatures are called hot spots. They dominate the reaction for shock initiation. The understanding of hot spot generation and their subsequent evolution has been limited by the inability to measure transients on small spatial ({approx} 1 {micro}m) and small temporal ({approx} 1 ns) scales in the harsh environment of a detonation. With the advances in computing power, it is natural to try and gain an understanding of hot-spot initiation with numerical experiments based on meso-scale simulations that resolve material heterogeneities and utilize realistic chemical reaction rates. However, to capture the underlying physics correctly, such high resolution simulations will require more than fast computers with a large amount of memory. Here we discuss some of the issues that need to be addressed. These include dissipative mechanisms that generate hot spots, accurate thermal propceties for the equations of state of the reactants and products, and controlling numerical entropy error from shock impedance mismatches at material interfaces. The later can generate artificial hot spots and lead to premature reaction. Eliminating numerical hot spots is critical for shock initiation simulations due to the positive feedback between the energy release from reaction and the hydrodynamic flow.

45 MILITARY TECHNOLOGY, WEAPONRY, AND NATIONAL DEF↗

Contention Modeling for Multithreaded Distributed Shared Memory Machines: The Cray XMT

Distributed Shared Memory (DSM) machines are a wide class of multi-processor computing systems where a large virtually-shared address space is mapped on a network of physically distributed memories. High memory latency and network contention are two of the main factors that limit performance scaling of such architectures. Modern high-performance computing DSM systems have evolved toward exploitation of massive hardware multi-threading and fine-grained memory hashing to tolerate irregular latencies, avoid network hot-spots and enable high scaling. In order to model the performance of such large-scale machines, parallel simulation has been proved to be a promising approach to achieve good accuracy in reasonable times. One of the most critical factors in solving the simulation speed-accuracy trade-off is network modeling. The Cray XMT is a massively multi-threaded supercomputing architecture that belongs to the DSM class, since it implements a globally-shared address space abstraction on top of a physically distributed memory substrate. In this paper, we discuss the development of a contention-aware network model intended to be integrated in a full-system XMT simulator. We start by measuring the effects of network contention in a 128-processor XMT machine and then investigate the trade-off that exists between simulation accuracy and speed, by comparing three network models which operate at different levels of accuracy. The comparison and model validation is performed by executing a string-matching algorithm on the full-system simulator and on the XMT, using three datasets that generate noticeably different contention patterns.

99 GENERAL AND MISCELLANEOUS↗