Search NASA⌕ Search

SEARCH · Search NASA

Results for “parallel systems”

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 505 records · Page 28

Optimizing Management of Persistent Data Structures in High-Performance Analytics

Large-scale data analytics workflows ingest massive input data into various data structures, including graphs and key-value datastores. These data structures undergo multiple transformations and computations and are typically reused in incremental and iterative analytics workflows. Persisting in-memory views of these data structures enables reusing them beyond the scope of a single program run while avoiding repetitive raw data ingestion overheads. Memory-mapped I/O enables persisting in-memory data structures without data serialization and deserialization overheads. However, memory-mapped I/O lacks the key feature of persisting consistent snapshots of these data structures for incremental ingestion and processing. The obstacles to efficient virtual memory snapshots using memory-mapped I/O include background writebacks outside the application’s control, and the significantly high storage footprint of such snapshots. To address these limitations, we present Privateer, a memory and storage management tool that enables storage-efficient virtual memory snapshotting while also optimizing snapshot I/O performance. Here, we integrated Privateer into Metall, a state-of-the-art persistent memory allocator for C++, and the Lightning Memory-Mapped Database (LMDB), a widely-used key-value datastore in data analytics and machine learning. Privateer optimized application performance by 1.22× when storing data structure snapshots to node-local storage, and up to 16.7× when storing snapshots to a parallel file system. Privateer also optimizes storage efficiency of incremental data structure snapshots by up to 11× using data deduplication and compression.

Computer science↗

Parallel transmit excitation at 1.5 T based on the minimization of a driving function for device heating

Purpose: To provide a rapid method to reduce the radiofrequency (RF) E-field coupling and consequent heating in long conductors in an interventional MRI (iMRI) setup. Methods: A driving function for device heating (W) was defined as the integration of the E-field along the direction of the wire and calculated through a quasistatic approximation. Based on this function, the phases of four independently controlled transmit channels were dynamically changed in a 1.5 T MRI scanner. During the different excitation configurations, the RF induced heating in a nitinol wire immersed in a saline phantom was measured by fiber-optic temperature sensing. Additionally, a minimization of W as a function of phase and amplitude values of the different channels and constrained by the homogeneity of the RF excitation field (B{sub 1}) over a region of interest was proposed and its results tested on the benchtop. To analyze the validity of the proposed method, using a model of the array and phantom setup tested in the scanner, RF fields and SAR maps were calculated through finite-difference time-domain (FDTD) simulations. In addition to phantom experiments, RF induced heating of an active guidewire inserted in a swine was also evaluated. Results: In the phantom experiment, heating at the tip of the device was reduced by 92% when replacing the body coil by an optimized parallel transmit excitation with same nominal flip angle. In the benchtop, up to 90% heating reduction was measured when implementing the constrained minimization algorithm with the additional degree of freedom given by independent amplitude control. The computation of the optimum phase and amplitude values was executed in just 12 s using a standard CPU. The results of the FDTD simulations showed similar trend of the local SAR at the tip of the wire and measured temperature as well as to a quadratic function of W, confirming the validity of the quasistatic approach for the presented problem at 64 MHz. Imaging and heating reduction of the guidewire were successfully performed in vivo with the proposed hardware and phase control. Conclusions: Phantom and in vivo data demonstrated that additional degrees of freedom in a parallel transmission system can be used to control RF induced heating in long conductors. A novel constrained optimization approach to reduce device heating was also presented that can be run in just few seconds and therefore could be added to an iMRI protocol to improve RF safety.

60 APPLIED LIFE SCIENCES↗

Interaction of carbon dioxide laser radiation with a nanotube array in the presence of a constant electric field

The dependence of the current density on the leading edge width of the alternating (high-frequency) field amplitude is studied at various constant (or unsteady) fields. The dependence of amplified microwaves in the two-millimeter range on a longitudinal coordinate is determined. The problem of submillimeter radiation generation in a system of parallel carbon nanotubes exposed to two-frequency carbon dioxide (CO{sub 2} laser) laser radiation in the presence of a constant (or unsteady) field is studied. The possibility of using freely oriented carbon nanotubes parallel to each other is shown.

36 MATERIALS SCIENCE↗

Linear chains of Ge/Si quantum dots grown on a prepatterned surface formed by ion irradiation

The growth of Ge nanoclusters on a prepatterned Si (100) surface formed by imprint lithography in combination with subsequent irradiation with Ge{sup +} ions is studied. The prepatterned surface presents a system of parallel 10-nm-wide grooves repeating with a period of 180 nm. Irradiation of the substrate was conducted at two temperatures, room temperature (cold irradiation) and 400°C (hot irradiation). It is shown that, during epitaxy (550–700°C), the residual radiation defects located in the bulk under the grooves suppress the nucleation of Ge nanoclusters in the grooves. In the case of prepatterned substrates, from which imperfect regions are completely removed, nanoclusters grow in the grooves.

36 MATERIALS SCIENCE↗

Optimizing I/O Performance of HPC Applications with Autotuning

Parallel Input output is an essential component of modern high-performance computing (HPC). Obtaining good I/O performance for a broad range of applications on diverse HPC platforms is a major challenge, in part, because of complex inter dependencies between I/O middleware and hardware. The parallel file system and I/O middleware layers all offer optimization parameters that can, in theory, result in better I/O performance. Unfortunately, the right combination of parameters is highly dependent on the application, HPC platform, problem size, and concurrency. Scientific application developers do not have the time or expertise to take on the substantial burden of identifying good parameters for each problem configuration. They resort to using system defaults, a choice that frequently results in poor I/O performance. We expect this problem to be compounded on exascale-class machines, which will likely have a deeper software stack with hierarchically arranged hardware resources. We present as a solution to this problem an autotuning system for optimizing I/O performance, I/O performance modeling, I/O tuning, and I/O patterns. We demonstrate the value of this framework across several HPC platforms and applications at scale.

Behzad, Babak↗

A massively parallel and scalable multi-CPU material point method

Harnessing the power of modern multi-GPU architectures, we present a massively parallel simulation system based on the Material Point Method (MPM) for simulating physical behaviors of materials undergoing complex topological changes, self-collision, and large deformations. Our system makes three critical contributions. First, we introduce a new particle data structure that promotes coalesced memory access patterns on the GPU and eliminates the need for complex atomic operations on the memory hierarchy when writing particle data to the grid. Second, we propose a kernel fusion approach using a new Grid-to-Particles-to-Grid (G2P2G) scheme, which efficiently reduces GPU kernel launches, improves latency, and significantly reduces the amount of global memory needed to store particle data. Finally, we introduce optimized algorithmic designs that allow for efficient sparse grids in a shared memory context, enabling us to best utilize modern multi-GPU computational platforms for hybrid Lagrangian-Eulerian computational patterns. We demonstrate the effectiveness of our method with extensive benchmarks, evaluations, and dynamic simulations with elastoplasticity, granular media, and fluid dynamics. In comparisons against an open-source and heavily optimized CPU-based MPM codebase [Fang et al. 2019] on an elastic sphere colliding scene with particle counts ranging from 5 to 40 million, our GPU MPM achieves over 100x per-time-step speedup on a workstation with an Intel 8086K CPU and a single Quadro P6000 GPU, exposing exciting possibilities for future MPM simulations in computer graphics and computational science. Moreover, compared to the state-of-the-art GPU MPM method [Hu et al. 2019a], we not only achieve 2x acceleration on a single GPU but our kernel fusion strategy and Array-of-Structs-of-Array (AoSoA) data structure design also generalizes to multi-GPU systems. Our multi-GPU MPM exhibits near-perfect weak and strong scaling with 4 GPUs, enabling performant and large-scale simulations on a 10243 grid with close to 100 million particles with less than 4 minutes per frame on a single 4-GPU workstation and 134 million particles with less than 1 minute per frame on an 8-GPU workstation.

Wang, Xinlei↗

Large-Scale Modeling of Epileptic Seizures: Scaling Properties of Two Parallel Neuronal Network Simulation Algorithms

Our limited understanding of the relationship between the behavior of individual neurons and large neuronal networks is an important limitation in current epilepsy research and may be one of the main causes of our inadequate ability to treat it. Addressing this problem directly via experiments is impossibly complex; thus, we have been developing and studying medium-large-scale simulations of detailed neuronal networks to guide us. Flexibility in the connection schemas and a complete description of the cortical tissue seem necessary for this purpose. In this paper we examine some of the basic issues encountered in these multiscale simulations. We have determined the detailed behavior of two such simulators on parallel computer systems. The observed memory and computation-time scaling behavior for a distributed memory implementation were very good over the range studied, both in terms of network sizes (2,000 to 400,000 neurons) and processor pool sizes (1 to 256 processors). Our simulations required between a few megabytes and about 150 gigabytes of RAM and lasted between a few minutes and about a week, well within the capability of most multinode clusters. Therefore, simulations of epileptic seizures on networks with millions of cells should be feasible on current supercomputers.

59 BASIC BIOLOGICAL SCIENCES↗

NUMA-Aware Thread Scheduling for Big Data Transfers over Terabits Network Infrastructure

The evergrowing trend of big data has led scientists to share and transfer the simulation and analytical data across the geodistributed research and computing facilities. However, the existing data transfer frameworks used for data sharing lack the capability to adopt the attributes of the underlying parallel file systems (PFS). LADS (Layout-Aware Data Scheduling) is an end-to-end data transfer tool optimized for terabit network using a layout-aware data scheduling via PFS. However, it does not consider the NUMA (Nonuniform Memory Access) architecture. In this paper, we propose a NUMA-aware thread and resource scheduling for optimized data transfer in terabit network. First, we propose distributed RMA buffers to reduce memory controller contention in CPU sockets and then schedule the threads based on CPU socket and NUMA nodes inside CPU socket to reduce memory access latency. We design and implement the proposed resource and thread scheduling in the existing LADS framework. Experimental results showed from 21.7% to 44% improvement with memory-level optimizations in the LADS framework as compared to the baseline without any optimization.

97 MATHEMATICS AND COMPUTING↗

nem_spread Ver. 5.10

Nem_spread reads it's input command file (default name nem_spread.inp), takes the named ExodusII geometry definition and spreads out the geometry (and optionally results) contained in that file out to a parallel disk system. The decomposition is taken from a scalar Nemesis load balance file generated by the companion utility nem_slice.

HENNIGAN, GARY↗

UPC++

UPC++ is a Partitioned Global Address Space (PGAS) extension for C++, with three main objectives: 1) to provide an object-oriented PGAS programming model in the context of the popular C++ language; 2) to add useful parallel programming idioms unavailable in UPC, such as asynchronous remote function invocation and multidimensional arrays, to support complex scientific applications; 3) to offer an easy on-ramp to PGAS programming through interoperability with other existing parallel programming systems (e.g., MPI, OpenMP, CUDA). UPC++ includes two main features beyond UPC: 1) One of the major limitations of UPC in the context of DOE application is the relatively weak notion of multidimensional arrays, which it inherited from C. To remedy this problem in UPC++, we included a multidimensional domain and array library based on that of Titanium. Titanium's library is similar to Chapel's dense and strided domains, as both were inspired by the dense and strided regions and arrays in ZPL. 2) Another feature provided by UPC++ but not by UPC is remote function invocation, which is inspired by Phalanx, X10, Habanero-C, and C++11. The user may start an asynchronous remote function invocation with the following syntax: "async(place)(function, args...);". UPC++ provides two programming constructs for specifying dynamic dependencies among tasks: 1) event-driven execution as in Phalanx and Habanero-C; 2) finish-async as in X10. We implemented UPC++ with a "compiler-free" approach using C++ templates and runtime libraries. The UPC++ front-end consists of a set of C++ header files that enables the C++ compiler to "translate" UPC++ features in a user program to runtime library calls. In particular, we use templates and operator overloading to customize the behavior of UPC++ types. This software package has been replaced with UPCC++ V 1.0

Amir Kamil, YiliZheng↗

Sirocco Storage Server v. pre-alpha 0.1

Sirocco is a parallel storage system under development, designed for write-intensive workloads on large-scale HPC platforms. It implements a keyvalue object store on top of a set of loosely federated storage servers that cooperate to ensure data integrity and performance. It includes support for a range of different types of storage transactions. This software release constitutes a conformant storage server, along with the client-side libraries to access the storage over a network.

Curry, MatthewL.↗

Livermore Big Artificial Neural Network Toolkit

LBANN is a toolkit that is designed to train artificial neural networks efficiently on high performance computing architectures. It is optimized to take advantages of key High Performance Computing features to accelerate neural network training. Specifically it is optimized for low-latency, high bandwidth interconnects, node-local NVRAM, node-local GPU accelerators, and high bandwidth parallel file systems. It is built on top of the open source Elemental distributed-memory dense and spars-direct linear algebra and optimization library that is released under the BSD license. The algorithms contained within LBANN are drawn from the academic literature and implemented to work within a distributed-memory framework.

Essen, BrianVan↗

TaskAMR

TaskAMR is a Cartesian Block Adaptive Mesh Refinement (AMR) code, written using the Regent/Legion parallel programming system, that solves Finite Volume calculations. We provide implementations for linear advection and gas dynamics.

Pietarila, Jonathan↗

PoliMOR

PoliMOR is a scalable, automated, and customizable policy engine framework for multi-tiered parallel file systems. It is composed of single-purpose agents that handle tasks such as gathering file metadata, making policy decisions, and then executing actions based on those policies. These agents are designed to communicate using distributed message queues, allowing the number of individual agents to be scaled up as needed. PoliMOR automates the data management tasks by precluding the need for admin intervention. The agents in PoliMOR can be customized to integrate any utilities/tools that perform tasks like metadata scanning and data placement management.

Brumgard, Christopher↗

Nodeman: A Node Management Tool For Hpc Clusters

NodeMan is a command line tool to manage nodes in an HPC cluster. At it's core, it is an extensible framework composed of bash scripting and GNU parallel. HPC System Administrator will find it useful in that it encapsulates desired functions and allows them to be assembled in a way familiar to administrators - through pipes. In fact, NodeMan functions can work with common command line tools as long as they use stdin/stdout. System Administrators can construct moderately complex logic and filtering on a compact command line that would normally require a substantial shell script. In the spirit of clush and pdsh, it is able to run commands remotely on nodes. Additionally, NodeMan is more flexible. For example, it can interact with IPMI and naturally processes node lists for orchestrating different tools. The library of useful pre-built functions is growing. System administrators can easily create new functions and make it their own.

Serr, ScottM↗

Attribution of North American Subseasonal Precipitation Prediction Skill

The skill of NOAA’s official monthly U.S. precipitation forecasts (issued in the middle of the prior month) has historically been low, having shown modest skill over the southern United States, but little or no skill over large portions of the central United States. The goal of this study is to explain the seasonal and regional variations of the North American subseasonal (weeks 3–6) precipitation skill, specifically the reasons for its successes and its limitations. The performances of multiple recent-generation model reforecasts over 1999–2015 in predicting precipitation are compared to uninitialized simulation skill using the atmospheric component of the forecast systems. This parallel analysis permits attribution of precipitation skill to two distinct sources: one due to slowly evolving ocean surface boundary states and the other to faster time-scale initial atmospheric weather states. A strong regionality and seasonality in precipitation forecast performance is shown to be analogous to skill patterns dictated by boundary forcing constraints alone. The correspondence is found to be especially high for the North American pattern of the maximum monthly skill that is achieved in the reforecast. The boundary forcing of most importance originates from tropical Pacific SST influences, especially those related to El Niño–Southern Oscillation. Furthermore, we discuss physical constraints that may limit monthly precipitation skill and interpret the performance of existing models in the context of plausible upper limits.

54 ENVIRONMENTAL SCIENCES↗

Demystifying asynchronous I/O Interference in HPC applications

With increasing complexity of HPC workflows, data management services need to perform expensive I/O operations asynchronously in the background, aiming to overlap the I/O with the application runtime. However, this may cause interference due to competition for resources: CPU, memory/network bandwidth. The advent of multi-core architectures has exacerbated this problem, as many I/O operations are issued concurrently, thereby competing not only with the application but also among themselves. Furthermore, the interference patterns can dynamically change as a response to variations in application behavior and I/O subsystems (e.g. multiple users sharing a parallel file system). Without a thorough understanding, I/O operations may perform suboptimally, potentially even worse than in the blocking case. To fill this gap, here we investigate the causes and consequences of interference due to asynchronous I/O on HPC systems. Specifically, we focus on multi-core CPUs and memory bandwidth, isolating the interference due to each resource. Then, we perform an in-depth study to explain the interplay and contention in a variety of resource sharing scenarios such as varying priority and number of background I/O threads and different I/O strategies: sendfile, read/write, mmap/write underlining trade-offs. The insights from this study are important both to enable guided optimizations of existing background I/O, as well as to open new opportunities to design advanced asynchronous I/O strategies.

97 MATHEMATICS AND COMPUTING↗

AI4IO: A suite of AI-based tools for IO-aware scheduling

Traditional workload managers do not have the capacity to consider how IO contention can increase job runtime and even cause entire resource allocations to be wasted. Whether from bursts of IO demand or parallel file systems (PFS) performance degradation, IO contention must be identified and addressed to ensure maximum performance. In this paper, we present AI4IO (AI for IO), a suite of tools using AI methods to prevent and mitigate performance losses due to IO contention. AI4IO enables existing workload managers to become IO-aware. Currently, AI4IO consists of two tools: PRIONN and CanarIO. PRIONN predicts IO contention and empowers schedulers to prevent it. CanarIO mitigates the impact of IO contention when it does occur. We measure the effectiveness of AI4IO when integrated into Flux, a next-generation scheduler, for both small- and large-scale IO-intensive job workloads. Our results show that integrating AI4IO into Flux improves the workload makespan up to 6.4%, which can account for more than 18,000 node-h of saved resources per week on a production cluster in our large-scale workload.

Wyatt, II, Michael R.↗