Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “tensor program generation”

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.

24 records · Page 2
All sourcesNASA NTRSDOE OSTI
DOE OSTI2017

ARM West Antarctic Radiation Experiment (AWARE) Fourier Transform Infrared Filter Sampling Systems Field Campaign Report

West Antarctica is one of the most rapidly warming regions in the world (Bromwich et al. 2013; Steig et al. 2009) and the warming can directly cause sea level rise (King et al. 2012). In situ measurements are needed to examine and quantify the meteorological conditions in order to precisely understand the effect of warm air and clouds on the West Antarctic Ice Shelf (WAIS). The U.S. Department of Energy (DOE) Atmospheric Radiation Measurement (ARM) West Antarctic Radiation Experiment (AWARE) is designed to collect measurements with advanced cloud and aerosol instrumentation at West Antarctica. The Fourier Transform Infrared (FTIR) filter sampling systems were deployed as part of the ARM West Antarctic Radiation Experiment (AWARE) campaign to measure organic aerosol characteristics at McMurdo Station. As a coastal station at the southern end of Ross Island, McMurdo Station has a meteorological relationship with the WAIS via circulation patterns in the Ross and Amundsen Seas. Thirteen months of continuous observations of aerosol particle composition extended from December 2016 to January 2017. Filter samples were collected at McMurdo on pre-scanned Teflon filters (Teflon, Pall Inc., 37mm diameter) behind PM1 sharp-cut cyclones (SCC2.229 PM1, BGI Inc). Samples were collected each week with approximately 50-90 m3 air filtered per sample. Samples were frozen and transported below 0°C to the Scripps laboratory for FTIR spectroscopy analysis. A Bruker Tensor 27 FTIR spectrometer with a deuterated triglycine sulfate (DTGS) detector (Bruker, Waltham, MA) was used to scan the filters both before and after sampling. An automated algorithm (Takahama et al. 2013) was applied to quantify the mass of the organic functional groups. Four groups (alkane, amine, alcohol, and carboxylic acid) had absorption peaks above the limits of quantification, and the sum of the mass of the four functional groups is the quantified organic mass (OM). Absorption peaks for other groups (organosulfate, organonitrate, aromatic, and alkene group) were identified and fit but they were below the limit of quantification for more than 90% of the samples and are excluded. Similar FTIR measurements has been conducted in dozens of field campaigns around the world, including arctic regions (Russell et al. 2011). Local particle emissions from airports, diesel generators, and local transportation affected all aerosol measurements at various times during this project. The ARM Aerosol Measurement Science Group is aware of the issue and is working to develop algorithms to flag the data. This influence is evident in the FTIR measurements as it is associated with specific chemical signatures. DOE and the National Science Foundation, which manages the U.S. Antarctic Program, supported this collaboration for AWARE.

ARM West Antarctic Radiation Experiment, Fourier T↗
DOE OSTI2023

MOOSE Reactor Module: An Open-Source Capability for Meshing Nuclear Reactor Geometries

The U.S. Department of Energy (DOE) Nuclear Energy Advanced Modeling and Simulation (NEAMS) program has developed numerous physics solvers utilizing the open-source Multiphysics Object-Oriented Simulation Environment (MOOSE) framework for multiphysics reactor analysis. These solvers require input finite element meshes representing the discretized spatial domain. Typically, reactor analysts turn to licensed tools for the creation of reactor geometry meshes. Recently, open-source functionality has been added to the MOOSE framework to mesh common reactor geometries and improve MOOSE-based nuclear reactor application user workflows. The new functionality is primarily contained in the new Reactor module of MOOSE and includes support for hexagonal pins, assemblies, and cores, extended Cartesian geometry support, options for modeling static and rotating control drums within a hexagonal assembly, core periphery triangulation, and automatic tagging of pin, assembly, plane, and depletion regions for easier post processing of physics results. A set of reactor geometry mesh builder objects further streamlines the construction of hexagonal and Cartesian cores and allows mapping of materials to regions during mesh generation. The meshes produced with the MOOSE Reactor module may be used directly within MOOSE-based applications or exported as Exodus II files for use in other finite element solvers. The tools have been demonstrated and verified using a variety of NEAMS physics solvers on a range of reactor applications, including a sodium-cooled fast reactor core analysis using Griffin, a fast reactor assembly thermal deformation analysis using MOOSE Tensor Mechanics, and a heat pipe–cooled microreactor coupled analysis using Griffin, Bison, and Sockeye. MOOSE’s Reactor module provides significant advantages compared to the use of external meshing tools when analyzing Cartesian and hexagonal reactor lattices using MOOSE-based applications: immediate accessibility (open-source) to the end user, low barrier to entry for new users, speed of mesh generation, volume preservation of meshed fuel pins, and simplification of analysis workflow when used in conjunction with MOOSE-based applications.

99 GENERAL AND MISCELLANEOUS↗
DOE OSTI2019

Future SDN-HPON Control Plane Architecture and Protocol for On-Demand Terabit End-to-End Extreme-Scale Science Applications

The UC Davis team conducted architecture, algorithm and experimental studies for next generation ultra-high-bandwidth optical networks in support of extreme-scale science applications. In particular, over the course of this 3-year project, the UC Davis team achieved the following main accomplishments: Designed a network architecture for automated multi-domain service provisioning: Designed information exchange schemes between autonomous systems (domains) Designed application programming interfaces to orchestrate resource reservation in data centers, high performance computation facilities, and communication networks belonging to different autonomous systems Experimentally assessed the multi-domain network architecture in a distributed field trial set-up connecting premises in three continents Designed a broker-based orchestrated multi-domain optical networking framework supporting observe-analyze-act cycle based cognitive inter-domain service provisioning: Designed an inter-domain operation mechanism, e.g., topology abstraction, for the broker plane and domain managers to collaboratively accomplish inter-domain service provisioning while respecting the autonomy of each administrative domain Designed an alien wavelength monitoring scheme and an artificial neural network based optical signal-to-noise ratio estimator for quality-of-transmission aware inter-domain service provisioning Developed a deep neural network based multi-domain traffic predictor and a cognitive routing and spectrum assignment algorithm to realize knowledge-based autonomous traffic engineering in multi-domain networks Implemented an optical signal-to-noise ratio estimator and traffic predictor with Google’s Tensor flow and assessed their performance using experimental and synthetic data Demonstrated the benefits of the proposed cognitive schemes with experiments ona two-domain elastic optical network testbed and evaluations on a six-domain simulation framework Built a two-domain software-defined networking control plane testbed using the ONOS platform and demonstrated the workflow of broker and domain manager cooperation, including hierarchical performance monitoring, topology abstraction and cognitive resource allocation. Designed self-learning cognitive service provisioning schemes using advanced machine learning technologies: Designed a self-taught anomaly detection framework for optical networks, including an unsupervised data clustering module for pattern recognition, a supervised data regression and classification module for online anomaly detection, and the system operation principle Experimentally validated the effectiveness of the self-taught anomaly detection framework with various failure/anomaly scenarios. Designed a self-learning autonomic routing, modulation and spectrum assignment (RMSA) provisioning framework for elastic optical networks with deep reinforcement learning. Designed effective modeling and training mechanisms for the self-learning RMSA framework Conducted extensive numerical simulations with various traffic patterns and topologies to verify the advantage of the self-learning RMSA framework.

42 ENGINEERING↗
DOE OSTI2022

Exploring the Use of Novel Spatial Accelerators in Scientific Applications

Driven by the need to find alternative accelerators which can viably replace GPUs in next-generation Supercomputing systems, this paper proposes a methodology to enable agile application/hardware co-design. The application-first methodology provides the ability to come up with design of accelerators while working with real-world workloads, available accelerators, and system software. The iterative design process targets a set of kernels in a workload for performance estimates that can prune the design space for later phases of detailed architectural evaluations. To this effect, in this paper, a novel data-parallel device model is introduced that simulates the latency of performance-sensitive operations in an accelerator including data transfers and kernel computation using multi-core CPUs. The use of off-the-shelf simulators, such as pre-RTL simulator Aladdin or multiple tools available for exploring the design of deep neural network accelerators (e.g., Timeloop) is demonstrated for evaluation of various accelerator designs using applications with realistic inputs. Examples of multiple device configurations that are instantiable in a system are explored to evaluate the performance benefit of deploying novel accelerators. The proposed device is integrated with a programming model and system software to potentially explore the impacts of high-level programming languages/compilers and low-level effects such as task scheduling on multiple accelerators. We analyze our methodology for a set of applications that represent high-performance computing (HPC) and graph analytics. The applications include a computational chemistry kernel realized using tensor contractions, triangle counting, GraphSAGE and Breadth-first Search. These applications include kernels such as dense matrix-dense matrix multiplication, sparse matrix-spare matrix multiplication, and sparse matrix-dense vector multiplication. Our results indicate potential performance benefits and insights for system design by including accelerators that realize these kernels along-side general purpose accelerators.

AI, codesign, Accelerated Computing, Modeling and ↗
DOE OSTI2022

Automatic Generation of High-Performance Convolution Kernels on ARM CPUs for Deep Learning

In this work, we present FastConv, a template-based code auto-generation open source library that can automatically generate high-performance deep learning convolution kernels of arbitrary matrices/tensors shapes. FastConv is based on the Winograd algorithm, which is reportedly the highest performing algorithm for the time-consuming convolution layers of convolutional neural networks. ARM CPUs cover a wide range designs and specifications, from embedded devices to HPC-grade CPUs. The leads to the dilemma of how to consistently optimize Winograd-based convolution solvers for convolution layers of different shapes. FastConv addresses this problem by using templates to auto-generate multiple shapes of tuned kernels variants suitable for skinny tall matrices. As a performance portable library, FastConv transparently searches for the best combination of kernel shapes, cache tiles, scheduling of loop orders, packing strategies, access patterns, and online/offline computations. Auto-tuning is used to search the parameter configuration space for the best performance for a given target architecture and problem size. The experiments with layer-wise evaluation on the VGG--16 model confirms a 1.25x performance gains is got by tuning the Winograd library. Integrated comparison results shows 1.02x to 1.40x, 1.14x to 2.17x, and 1.22x and 2.48x speedup is achieved over NNPACK, Arm NN, and FeatherCNN on the Kunpeng 920 beside few cases. Furthermore, problem size performance portability experiments with various convolution shapes shows that FastConv achieves 1.2x to 1.7x speedup and 2x to 22x speedup over NNPACK and ARM NN inference engine using Winograd on Kunpeng 920 . CPU performance portability evaluation on the VGG--16 show an average speedup over NNPACK of 1.42x, 1.21x, 1.26x, 1.37x, 2.26x, and 11.02x is observed on Kunpeng 920, Snapdragon 835, 855, 888, Apple M1, and AWS Graviton2, respectively.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2016

Lattice QCD Calculation of Nucleon Structure

It is emphasized in the 2015 NSAC Long Range Plan that "understanding the structure of hadrons in terms of QCD's quarks and gluons is one of the central goals of modern nuclear physics." Over the last three decades, lattice QCD has developed into a powerful tool for ab initio calculations of strong-interaction physics. Up until now, it is the only theoretical approach to solving QCD with controlled statistical and systematic errors. Since 1985, we have proposed and carried out first-principles calculations of nucleon structure and hadron spectroscopy using lattice QCD which entails both algorithmic development and large-scale computer simulation. We started out by calculating the nucleon form factors -- electromagnetic, axial-vector, πNN, and scalar form factors, the quark spin contribution to the proton spin, the strangeness magnetic moment, the quark orbital angular momentum, the quark momentum fraction, and the quark and glue decomposition of the proton momentum and angular momentum. The first round of calculations were done with Wilson fermions in the `quenched' approximation where the dynamical effects of the quarks in the sea are not taken into account in the Monte Carlo simulation to generate the background gauge configurations. Beginning in 2000, we have started implementing the overlap fermion formulation into the spectroscopy and structure calculations. This is mainly because the overlap fermion honors chiral symmetry as in the continuum. It is going to be more and more important to take the symmetry into account as the simulations move closer to the physical point where the u and d quark masses are as light as a few MeV only. We began with lattices which have quark masses in the sea corresponding to a pion mass at ~ 300 MeV and obtained the strange form factors, charm and strange quark masses, the charmonium spectrum and the D s meson decay constant f D s , the strangeness and charmness, the meson mass decomposition and the strange quark spin from the anomalous Ward identity. Recently, we have started to include multiple lattices with different lattice spacings and different volumes including large lattices at the physical pion mass point. We are getting quite close to being able to calculate the hadron structure at the physical point and to do the continuum and large volume extrapolations, which is our ultimate aim. We have now finished several projects which have included these systematic corrections. They include the leptonic decay width of the ρ, the πN sigma and strange sigma terms, and the strange quark magnetic moment. Over the years, we have also studied hadron spectroscopy with lattice calculations and in phenomenology. These include Roper resonance, pentaquark state, charmonium spectrum, glueballs, scalar mesons a 0 (1450) and σ(600) and other scalar mesons, and the 1 -+ meson. In addition, we have employed the canonical approach to explore the first-order phase transition and the critical point at finite density and finite temperature. We have also discovered a new parton degree of freedom -- the connected sea partons, from the path-integral formulation of the hadronic tensor, which explains the experimentally observed Gottfried sum rule violation. Combining experimental result on the strange parton distribution, the CT10 global fitting results of the total u and d anti-partons and the lattice result of the ratio of the momentum fraction of the strange vs that of u or d in the disconnected insertion, we have shown that the connected sea partons can be isolated. In this final technical report, we shall present a few representative highlights that have been achieved in the project.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗