Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “task-based algorithms”

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.

50 records · Page 3

Large Scale Parallel Solution of Incompressible Flow Problems Using Uintah and Hypre

The Uintah Software framework was developed to provide an environment for solving fluid-structure interaction problems on structured adaptive grids on large-scale, long-running, data-intensive problems. Uintah uses a combination of fluid-flow solvers and particle-based methods for solids together with a novel asynchronous task-based approach with fully automated load balancing. As Uintah is often used to solve incompressible flow problems in combustion applications it is important to have a scalable linear solver. While there are many such solvers available, the scalability of those codes varies greatly. The hypre software offers a range of solvers and pre-conditioners for different types of grids. The weak scalability of Uintah and hypre is addressed for particular examples of both packages when applied to a number of incompressible flow problems. After careful software engineering to reduce startup costs, much better than expected weak scalability is seen for up to 100K cores on NSFs Kraken architecture and up to260K cpu cores, on DOEs new Titan machine. The scalability is found to depend in a crtitical way on the choice of algorithm used by hypre for a realistic application problem.

Computer Science↗

ExaTN

A challenging scientific problem that will benefit from exascale computing is predictive simulation of two- and three-dimensional quantum many-body Hamiltonians in condensed matter physics, together with the related strong entanglement problems in quantum chemistry and quantum computing. Efficient numerical simulations will require new insights into wavefunction compression algorithms that scale on extremely heterogeneous HPC architectures. To address this need, we propose to develop a scalable math library of parallel multi-linear algebraic numerical primitives for hierarchical tensor compression. Our library will subsequently be used for building an efficient, performance-portable HPC framework for simulating strongly correlated quantum systems. In particular, our will will be the first to deliver a massively parallel implementation of the multiscale entanglement renormalization ansatz (MERA), a hierarchical tensor decomposition quantum scheme capable of expressing local expectation values of strongly entangled higher-dimensional quantum systems efficiently. To address the extreme complexity and high computational cost of such schemes, we will automate their parallel implementation by delivering an integrated software framework, ExaTN (Exascale Tensor Networks), that will process hierarchical tensor representations on exascale HPC systems via an asymchronous task-based programming model. As a specific application, our framework will provide scientists with a capability of simulating strongly entangled quantum systems, thus positioning ORNL as a strategic leader in modeling of strongly correlated materials.

McCaskey, AlexanderJ↗

ATDM LANL FleCSI: Topology and Execution Framework

FleCSI is a compile-time configurable C++ framework designed to support multi-physics application development. As such, FleCSI attempts to provide a very general set of infrastructure design patterns that can be specialized and extended to suit the needs of a broad variety of solver and data requirements. This means that FleCSI is potentially useful to many different ECP projects. Current support includes multidimensional mesh topology, mesh geometry, and mesh adjacency information, n-dimensional hashed-tree data structures, graph partitioning interfaces, and dependency closures (to identify data dependencies between distributed-memory address spaces). FleCSI introduces a functional programming model with control, execution, and data abstractions that are consistent with state-of-the-art task-based runtimes such as Legion and Charm++. The model also provides support for fine-grained, data-parallel execution with backend support for runtimes such as OpenMP and C++17. The FleCSI abstraction layer provides the developer with insulation from the underlying runtimes, while allowing support for multiple runtime systems, including conventional models like asynchronous MPI. The intent is to give developers a concrete set of user-friendly programming tools that can be used now, while allowing flexibility in choosing runtime implementations and optimizations that can be applied to architectures and runtimes that arise in the future. This project is essential to the ECP Ristra Next-Generation Code project, part of ASC ATDM, because it provides a hierarchically parallel programming model that is consistent with the design of modern system architectures, but which allows for the straightforward expression of algorithmic parallelism in a portably performant manner.

97 MATHEMATICS AND COMPUTING↗

SU-F-J-178: A Computer Simulation Model Observer for Task-Based Image Quality Assessment in Radiation Therapy

Purpose: Traditionally, image quality in radiation therapy is assessed subjectively or by utilizing physically-based metrics. Some model observers exist for task-based medical image quality assessment, but almost exclusively for diagnostic imaging tasks. As opposed to disease diagnosis, the task for image observers in radiation therapy is to utilize the available images to design and deliver a radiation dose which maximizes patient disease control while minimizing normal tissue damage. The purpose of this study was to design and implement a new computer simulation model observer to enable task-based image quality assessment in radiation therapy. Methods: A modular computer simulation framework was developed to resemble the radiotherapy observer by simulating an end-to-end radiation therapy treatment. Given images and the ground-truth organ boundaries from a numerical phantom as inputs, the framework simulates an external beam radiation therapy treatment and quantifies patient treatment outcomes using the previously defined therapeutic operating characteristic (TOC) curve. As a preliminary demonstration, TOC curves were calculated for various CT acquisition and reconstruction parameters, with the goal of assessing and optimizing simulation CT image quality for radiation therapy. Sources of randomness and bias within the system were analyzed. Results: The relationship between CT imaging dose and patient treatment outcome was objectively quantified in terms of a singular value, the area under the TOC (AUTOC) curve. The AUTOC decreases more rapidly for low-dose imaging protocols. AUTOC variation introduced by the dose optimization algorithm was approximately 0.02%, at the 95% confidence interval. Conclusion: A model observer has been developed and implemented to assess image quality based on radiation therapy treatment efficacy. It enables objective determination of appropriate imaging parameter values (e.g. imaging dose). Framework flexibility allows for incorporation of additional modules to include any aspect of the treatment process, and therefore has great potential for both assessment and optimization within radiation therapy.

60 APPLIED LIFE SCIENCES↗

IRIS Reimagined: Advancements in Intelligent Runtime System for Task-Based Programming

Task-based programming models are gaining traction in scientific computing. IRIS is a portable runtime system that exploits multiple heterogeneous programming systems and can discover available resources and manage multiple diverse programming systems (e.g., CUDA, Hexagon, HIP, Level Zero, OpenCL, and OpenMP) simultaneously. It accounts for the constraints of task dependencies and provides customizable scheduling policies to map those tasks to heterogeneous devices. In this paper, we present new capabilities added to IRIS to improve its portability for heterogeneous programming, build-friendliness, and performance efficiency. The new additions include vendor-specific kernel support, a runtime system with a foreign function interface to eliminate writing wrapper or boilerplate code for heterogeneous kernels, an easy-to-use and configurable CMake-based build environment, automatic and efficient data transfers and orchestration, and the Hunter and DAGGER toolchains to evaluate IRIS’s task scheduling algorithms.

Miniskar, Narasinga Rao↗

A Block-Based Triangle Counting Algorithm on Heterogeneous Environments

Triangle counting is a fundamental building block in graph algorithms. In this article, we propose a block-based triangle counting algorithm to reduce data movement during both sequential and parallel execution. Our block-based formulation makes the algorithm naturally suitable for heterogeneous architectures. The problem of partitioning the adjacency matrix of a graph is well-studied. Our task decomposition goes one step further: it partitions the set of triangles in the graph. By streaming these small tasks to compute resources, we can solve problems that do not fit on a device. We demonstrate the effectiveness of our approach by providing an implementation on a compute node with multiple sockets, cores and GPUs. The current state-of-the-art in triangle enumeration processes the Friendster graph in 2.1 seconds, not including data copy time between CPU and GPU. Using that metric, our approach is 20 percent faster. When copy times are included, our algorithm takes 3.2 seconds. This is 5.6 times faster than the fastest published CPU-only time.

97 MATHEMATICS AND COMPUTING↗

Assessment of volumetric noise and resolution performance for linear and nonlinear CT reconstruction methods

Purpose: For nonlinear iterative image reconstructions (IR), the computed tomography (CT) noise and resolution properties can depend on the specific imaging conditions, such as lesion contrast and image noise level. Therefore, it is imperative to develop a reliable method to measure the noise and resolution properties under clinically relevant conditions. This study aimed to develop a robust methodology to measure the three-dimensional CT noise and resolution properties under such conditions and to provide guidelines to achieve desirable levels of accuracy and precision. Methods: The methodology was developed based on a previously reported CT image quality phantom. In this methodology, CT noise properties are measured in the uniform region of the phantom in terms of a task-based 3D noise-power spectrum (NPS{sub task}). The in-plane resolution properties are measured in terms of the task transfer function (TTF) by applying a radial edge technique to the rod inserts in the phantom. The z-direction resolution properties are measured from a supplemental phantom, also in terms of the TTF. To account for the possible nonlinearity of IR, the NPS{sub task} is measured with respect to the noise magnitude, and the TTF with respect to noise magnitude and edge contrast. To determine the accuracy and precision of the methodology, images of known noise and resolution properties were simulated. The NPS{sub task} and TTF were measured on the simulated images and compared to the truth, with criteria established to achieve NPS{sub task} and TTF measurements with <10% error. To demonstrate the utility of this methodology, measurements were performed on a commercial CT system using five dose levels, two slice thicknesses, and three reconstruction algorithms (filtered backprojection, FBP; iterative reconstruction in imaging space, IRIS; and sinogram affirmed iterative reconstruction with strengths of 5, SAFIRE5). Results: To achieve NPS{sub task} measurements with <10% error, the number of regions of interest needed to be greater than 65. To achieve TTF measurements with <10% error, the contrast-to-noise ratio of the edge needed to be ≥15, achievable by averaging multiple slices across the same edge. The NPS{sub task} measured on a commercial CT system showed IR's reduced noise (IRIS, 30% and SAFIRE5, 55%) and “waxier” texture (peak frequencies: FBP, 0.25 mm{sup −1}; IRIS, 0.23 mm{sup −1}; and SAFIRE5, 0.16 mm{sup −1}). The TTF measured within the axial plane showed improved in-plane resolution with SAFIRE5 at the TTF 50% frequency, f{sub 50} (FBP, 0.36–0.41 mm{sup −1}; SAFIRE5, 0.37–0.46 mm{sup −1}). The TTF measured along the axial direction showed improved z-direction resolution with thinner slice thickness (f{sub 50}: 0.6 mm, 0.35–0.79 mm{sup −1}; 1.5 mm, 0.22–0.3 mm{sup −1}) and with SAFIRE5 (f{sub 50}: FBP, 0.35–0.52 mm{sup −1}; SAFIRE5, 0.42–0.79 mm{sup −1}). Both in-plane and z-direction resolution of SAFIRE5 showed strong dependency on contrast, reflecting SAFIRE5's nonlinearity. Conclusions: A methodology was developed to measure three-dimensional CT noise and resolution properties for iterative reconstruction, especially at challenging measurement conditions with low contrast and high image noise. The methodology also demonstrated its utility for evaluating commercial CT systems.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

WE-EF-207-01: FEATURED PRESENTATION and BEST IN PHYSICS (IMAGING): Task-Driven Imaging for Cone-Beam CT in Interventional Guidance

Purpose: This work introduces a task-driven imaging framework that utilizes a patient-specific anatomical model, mathematical definition of the imaging task, and a model of the imaging system to prospectively design acquisition and reconstruction techniques that maximize task-based imaging performance. Utility of the framework is demonstrated in the joint optimization of tube current modulation and view-dependent reconstruction kernel in filtered-backprojection reconstruction and non-circular orbit design in model-based reconstruction. Methods: The system model is based on a cascaded systems analysis of cone-beam CT capable of predicting the spatially varying noise and resolution characteristics as a function of the anatomical model and a wide range of imaging parameters. Detectability index for a non-prewhitening observer model is used as the objective function in a task-driven optimization. The combination of tube current and reconstruction kernel modulation profiles were identified through an alternating optimization algorithm where tube current was updated analytically followed by a gradient-based optimization of reconstruction kernel. The non-circular orbit is first parameterized as a linear combination of bases functions and the coefficients were then optimized using an evolutionary algorithm. The task-driven strategy was compared with conventional acquisitions without modulation, using automatic exposure control, and in a circular orbit. Results: The task-driven strategy outperformed conventional techniques in all tasks investigated, improving the detectability of a spherical lesion detection task by an average of 50% in the interior of a pelvis phantom. The non-circular orbit design successfully mitigated photon starvation effects arising from a dense embolization coil in a head phantom, improving the conspicuity of an intracranial hemorrhage proximal to the coil. Conclusion: The task-driven imaging framework leverages a knowledge of the imaging task within a patient-specific anatomical model to optimize image acquisition and reconstruction techniques, thereby improving imaging performance beyond that achievable with conventional approaches. 2R01-CA-112163; R01-EB-017226; U01-EB-018758; Siemens Healthcare (Forcheim, Germany)

60 APPLIED LIFE SCIENCES↗

TH-CD-207B-09: Task-Driven Fluence Field Modulation Design for Model-Based Iterative Reconstruction in CT

Purpose: There has been increasing interest in integrating fluence field modulation (FFM) devices with diagnostic CT scanners for dose reduction purposes. Conventional FFM strategies, however, are often either based on heuristics or the analysis of filtered-backprojection (FBP) performance. This work investigates a prospective task-driven optimization of FFM for model-based iterative reconstruction (MBIR) in order to improve imaging performance at the same total dose as conventional strategies. Methods: The task-driven optimization framework utilizes an ultra-low dose 3D scout as a patient-specific anatomical model and a mathematical formation of the imaging task. The MBIR method investigated is quadratically penalized-likelihood reconstruction. The FFM objective function uses detectability index, d’, computed as a function of the predicted spatial resolution and noise in the image. To optimize performance throughout the object, a maxi-min objective was adopted where the minimum d’ over multiple locations is maximized. To reduce the dimensionality of the problem, FFM is parameterized as a linear combination of 2D Gaussian basis functions over horizontal detector pixels and projection angles. The coefficients of these bases are found using the covariance matrix adaptation evolution strategy (CMA-ES) algorithm. The task-driven design was compared with three other strategies proposed for FBP reconstruction for a calcification cluster discrimination task in an abdomen phantom. Results: The task-driven optimization yielded FFM that was significantly different from those designed for FBP. Comparing all four strategies, the task-based design achieved the highest minimum d’ with an 8–48% improvement, consistent with the maxi-min objective. In addition, d’ was improved to a greater extent over a larger area within the entire phantom. Conclusion: Results from this investigation suggests the need to re-evaluate conventional FFM strategies for MBIR. The task-based optimization framework provides a promising approach that maximizes imaging performance under the same total dose constraint.

60 APPLIED LIFE SCIENCES↗

A Comparative Critical Analysis of Modern Task-Parallel Runtimes

The rise in node-level parallelism has increased interest in task-based parallel runtimes for a wide array of application areas. Applications have a wide variety of task spawning patterns which frequently change during the course of application execution, based on the algorithm or solver kernel in use. Task scheduling and load balance regimes, however, are often highly optimized for specific patterns. This paper uses four basic task spawning patterns to quantify the impact of specific scheduling policy decisions on execution time. We compare the behavior of six publicly available tasking runtimes: Intel Cilk, Intel Threading Building Blocks (TBB), Intel OpenMP, GCC OpenMP, Qthreads, and High Performance ParalleX (HPX). With the exception of Qthreads, the runtimes prove to have schedulers that are highly sensitive to application structure. No runtime is able to provide the best performance in all cases, and those that do provide the best performance in some cases, unfortunately, provide extremely poor performance when application structure does not match the scheduler's assumptions.

Wheeler, Kyle Bruce↗

Task Parallel Incomplete Cholesky Factorization using 2D Partitioned-Block Layout

We introduce a task-parallel algorithm for sparse incomplete Cholesky factorization that utilizes a 2D sparse partitioned-block layout of a matrix. Our factorization algorithm follows the idea of algorithms-by-blocks by using the block layout. The algorithm-byblocks approach induces a task graph for the factorization. These tasks are inter-related to each other through their data dependences in the factorization algorithm. To process the tasks on various manycore architectures in a portable manner, we also present a portable tasking API that incorporates different tasking backends and device-specific features using an open-source framework for manycore platforms i.e., Kokkos. A performance evaluation is presented on both Intel Sandybridge and Xeon Phi platforms for matrices from the University of Florida sparse matrix collection to illustrate merits of the proposed task-based factorization. Experimental results demonstrate that our task-parallel implementation delivers about 26.6x speedup (geometric mean) over single-threaded incomplete Choleskyby- blocks and 19.2x speedup over serial Cholesky performance which does not carry tasking overhead using 56 threads on the Intel Xeon Phi processor for sparse matrices arising from various application problems.

2D layout↗

A Block-Based Triangle Counting Algorithm on Heterogeneous Environments

Triangle counting is a fundamental building block in graph algorithms. In this paper, we propose a block-based triangle counting algorithm to reduce data movement during both sequential and parallel execution. Our block-based formulation makes the algorithm naturally suitable for heterogeneous architectures. The problem of partitioning the adjacency matrix of a graph is well-studied. Our task decomposition goes one step further: it partitions the set of triangles in the graph. By streaming these small tasks to compute resources, we can solve problems that do not fit on a device. We demonstrate the effectiveness of our approach by providing an implementation on a compute node with multiple sockets, cores and GPUs. The current state-of-the-art in triangle enumeration processes the Friendster graph in 2.1 seconds, not including data copy time between CPU and GPU. Using that metric, our approach is 20 percent faster. When copy times are included, our algorithm takes 3.2 seconds. This is 5.6 times faster than the fastest published CPU-only time.

97 MATHEMATICS AND COMPUTING↗

O(N) Quantum Molecular Dynamics (QMD) Algorithms, Code Requirements, Release Notes, and Library Deployment

The Exascale Computing Project (ECP) Co-design Center for Particle Applications (CoPA) is addressing the challenges for particle applications to run on upcoming exascale computing architectures. This includes short- and long-range molecular dynamics (MD), particle-in-cell (PIC), and quantum molecular dynamics (QMD). This document addresses $\mathcal{O}(N)$ QMD. The goal of this CoPA team is to develop $\mathcal{O}(N)$ linear scaling methods for QMD optimized for new and next generation architectures. This will provide the capability to simulate larger systems and run for longer timeframes than is possible today. Focus is initially on the density matrix construction. This requires efficient matrix formats and linear algebra operations for distributed architectures. A significant reduction in the wallclock time for an MD timestep is necessary to achieve practical large scale QMD. Equally challenging is to develop scalable distributed implementations. This requires the exploration of efficient data formats (dense, sparse, etc.), data decomposition and distribution across compute nodes using block- and graph-based methods. Traditional programming models such as MPI/OpenMP, hybrid MPI/OpenMP/Accelerator, as well as asynchronous task-based models will be explored to achieve maximum parallelism. Parallelism at the level of a matrix operation, solver, and/or application needs to be explored to gain the most benefit on new and emerging architectures. This report serves as a “living document” of the additions and enhancements throughout the ECP CoPA project for $\mathcal{O}(N)$ QMD. The following sections address the applications we are working with, the quantum chemistry libraries (BML and PROGRESS), and our co-design exploration through the use of a proxy application.

36 MATERIALS SCIENCE↗

An Asynchronous Task-based Fan-Both Sparse Cholesky Solver

Systems of linear equations arise at the heart of many scientific and engineering applications. Many of these linear systems are sparse; i.e., most of the elements in the coefficient matrix are zero. Direct methods based on matrix factorizations are sometimes needed to ensure accurate solutions. For example, accurate solution of sparse linear systems is needed in shift-invert Lanczos to compute interior eigenvalues. The performance and resource usage of sparse matrix factorizations are critical to time-to-solution and maximum problem size solvable on a given platform. In many applications, the coefficient matrices are symmetric, and exploiting symmetry will reduce both the amount of work and storage cost required for factorization. When the factorization is performed on large-scale distributed memory platforms, communication cost is critical to the performance of the algorithm. At the same time, network topologies have become increasingly complex, so that modern platforms exhibit a high level of performance variability. This makes scheduling of computations an intricate and performance-critical task. In this paper, we investigate the use of an asynchronous task paradigm, one-sided communication and dynamic scheduling in implementing sparse Cholesky factorization (symPACK) on large-scale distributed memory platforms. Our solver symPACK relies on efficient and flexible communication primitives provided by the UPC++ library. Performance evaluation shows good scalability and that symPACK outperforms state-of-the-art parallel distributed memory factorization packages, validating our approach on practical cases.

Jacquelin, Mathias↗