Search NASA⌕ Search

SEARCH · Search NASA

Results for “Unstructured data”

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 451 records · Page 25

Scalable multilevel Monte Carlo methods exploiting parallel redistribution on coarse levels

Here, we study an element agglomeration coarsening strategy that requires data redistribution at coarse levels when the number of coarse elements becomes smaller than the number of MPI processes used on the finest level. The overall procedure generates coarse elements (general unstructured unions of fine grid elements) within the framework of element-based algebraic multigrid methods (or AMGe) studied previously. The AMGe-generated coarse spaces have the ability to exhibit approximation properties of the same order as the fine-level spaces since by construction they contain the piecewise polynomials of the same order as on the fine level. These approximation properties are key for the successful use of AMGe in multilevel solvers for nonlinear partial differential equations as well as for multilevel Monte Carlo (MLMC) simulations. The ability to coarsen without being constrained by the number of MPI processes, as described in the present paper, allows to improve the scalability of these solvers as well as the overall MLMC method. The paper illustrates this latter fact with detailed scalability study of MLMC simulations applied to model Darcy equations with a stochastic log-normal permeability field.

AMGe↗

An assessment of the adaptive unstructured tetrahedral grid, Euler Flow Solver Code FELISA

A three-dimensional solution-adaptive Euler flow solver for unstructured tetrahedral meshes is assessed, and the accuracy and efficiency of the method for predicting sonic boom pressure signatures about simple generic models are demonstrated. Comparison of computational and wind tunnel data and enhancement of numerical solutions by means of grid adaptivity are discussed. The mesh generation is based on the advancing front technique. The FELISA code consists of two solvers, the Taylor-Galerkin and the Runge-Kutta-Galerkin schemes, both of which are spacially discretized by the usual Galerkin weighted residual finite-element methods but with different explicit time-marching schemes to steady state. The solution-adaptive grid procedure is based on either remeshing or mesh refinement techniques. An alternative geometry adaptive procedure is also incorporated.

Djomehri, M. Jahed↗

Computational Analyses in Support of Sub-scale Diffuser Testing for the A-3 Facility: Steady Predictions - Part 1

Simulation technology can play an important role in rocket engine test facility design and development by assessing risks, providing analysis of dynamic pressure and thermal loads, identifying failure modes and predicting anomalous behavior of critical systems. Advanced numerical tools assume greater significance in supporting testing and design of high altitude testing facilities and plume induced testing environments of high thrust engines because of the greater inter-dependence and synergy in the functioning of the different sub-systems. This is especially true for facilities such as the proposed A-3 facility at NASA SSC because of a challenging operating envelope linked to variable throttle conditions at relatively low chamber pressures. Facility designs in this case will require a complex network of diffuser ducts, steam ejector trains, fast operating valves, cooling water systems and flow diverters that need to be characterized for steady state performance. In this paper, we will demonstrate with the use of CFD analyses s advanced capability to evaluate supersonic diffuser and steam ejector performance in a sub-scale A-3 facility at NASA Stennis Space Center (SSC) where extensive testing was performed. Furthermore, the focus in this paper relates to modeling of critical sub-systems and components used in facilities such as the A-3 facility. The work here will address deficiencies in empirical models and current CFD analyses that are used for design of supersonic diffusers/turning vanes/ejectors as well as analyses for confined plumes and venting processes. The primary areas that will be addressed are: (1) supersonic diffuser performance including analyses of thermal loads (2) accurate shock capturing in the diffuser duct; (3) effect of turning duct on the performance of the facility (4) prediction of mass flow rates and performance classification for steam ejectors (5) comparisons with test data from sub-scale diffuser testing and assessment of confidence levels in CFD based flowpath modeling of the facility. The analyses tools used here expand on the multi-element unstructured CFD which has been tailored and validated for impingement dynamics of dry plumes, complex valve/feed systems, and high pressure propellant delivery systems used in engine and component test stands at NASA SSC. The analyses performed in the evaluation of the sub-scale diffuser facility explored several important factors that influence modeling and understanding of facility operation such as (a) importance of modeling the facility with Real Gas approximation, (b) approximating the cluster of steam ejector nozzles as a single annular nozzle, (c) existence of mixed subsonic/supersonic flow downstream of the turning duct, and (d) inadequacy of two-equation turbulence models in predicting the correct pressurization in the turning duct and expansion of the second stage steam ejectors. The procedure used for modeling the facility was as follows: (i) The engine, test cell and first stage ejectors were simulated with an axisymmetric approximation (ii) the turning duct, second stage ejectors and the piping downstream of the second stage ejectors were analyzed with a three-dimensional simulation utilizing a half-plane symmetry approximation. The solution i.e. primitive variables such as pressure, velocity components, temperature and turbulence quantities were passed from the first computational domain and specified as a supersonic boundary condition for the second simulation. (iii) The third domain comprised of the exit diffuser and the region in the vicinity of the facility (primary included to get the correct shock structure at the exit of the facility and entrainment characteristics). The first set of simulations comprising the engine, test cell and first stage ejectors was carried out both as a turbulent real gas calculation as well as a turbulent perfect gas calculation. A comparison for the two cases (Real Turbulent and Perfect gas turbulent) of the Ma Number distribution and temperature distributions are shown in Figures 1 and 2 respectively. The Mach Number distribution shows small yet distinct differences between the two cases such as locations of shocks/shock reflections and a slightly different impingement point on the wall of the diffuser from the expansion at the exit of the nozzle. Similarly the temperature distribution indicates different flow recirculation patterns in the test cell. Both cases capture all the essential flow phenomena such as the shock-boundary layer interaction, plume expansion, expansion of the first stage ejectors, mixing between the engine plume and the first stage ejector flow and pressurization due to the first stage ejectors. The final paper will discuss thermal loads on the walls of the diffuser and cooling mechanisms investigated.

Allgood, Daniel C.↗

Computational Analyses in Support of Sub-scale Diffuser Testing for the A-3 Facility: Steady Predictions - Part 1

Simulation technology can play an important role in rocket engine test facility design and development by assessing risks, providing analysis of dynamic pressure and thermal loads, identifying failure modes and predicting anomalous behavior of critical systems. Advanced numerical tools assume greater significance in supporting testing and design of high altitude testing facilities and plume induced testing environments of high thrust engines because of the greater inter-dependence and synergy in the functioning of the different sub-systems. This is especially true for facilities such as the proposed A-3 facility at NASA SSC because of a challenging operating envelope linked to variable throttle conditions at relatively low chamber pressures. Facility designs in this case will require a complex network of diffuser ducts, steam ejector trains, fast operating valves, cooling water systems and flow diverters that need to be characterized for steady state performance. In this paper, we will demonstrate with the use of CFD analyses s advanced capability to evaluate supersonic diffuser and steam ejector performance in a sub-scale A-3 facility at NASA Stennis Space Center (SSC) where extensive testing was performed. Furthermore, the focus in this paper relates to modeling of critical sub-systems and components used in facilities such as the A-3 facility. The work here will address deficiencies in empirical models and current CFD analyses that are used for design of supersonic diffusers/turning vanes/ejectors as well as analyses for confined plumes and venting processes. The primary areas that will be addressed are: (1) supersonic diffuser performance including analyses of thermal loads (2) accurate shock capturing in the diffuser duct; (3) effect of turning duct on the performance of the facility (4) prediction of mass flow rates and performance classification for steam ejectors (5) comparisons with test data from sub-scale diffuser testing and assessment of confidence levels in CFD based flowpath modeling of the facility. The analyses tools used here expand on the multi-element unstructured CFD which has been tailored and validated for impingement dynamics of dry plumes, complex valve/feed systems, and high pressure propellant delivery systems used in engine and component test stands at NASA SSC. The analyses performed in the evaluation of the sub-scale diffuser facility explored several important factors that influence modeling and understanding of facility operation such as (a) importance of modeling the facility with Real Gas approximation, (b) approximating the cluster of steam ejector nozzles as a single annular nozzle, (c) existence of mixed subsonic/supersonic flow downstream of the turning duct, and (d) inadequacy of two-equation turbulence models in predicting the correct pressurization in the turning duct and expansion of the second stage steam ejectors. The procedure used for modeling the facility was as follows: (i) The engine, test cell and first stage ejectors were simulated with an axisymmetric approximation (ii) the turning duct, second stage ejectors and the piping downstream of the second stage ejectors were analyzed with a three-dimensional simulation utilizing a half-plane symmetry approximation. The solution i.e. primitive variables such as pressure, velocity components, temperature and turbulence quantities were passed from the first computational domain and specified as a supersonic boundary condition for the second simulation. (iii) The third domain comprised of the exit diffuser and the region in the vicinity of the facility (primary included to get the correct shock structure at the exit of the facility and entrainment characteristics). The first set of simulations comprising the engine, test cell and first stage ejectors was carried out both as a turbulent real gas calculation as well as a turbulent perfect gas calculation. A comparison for the two cases (Real Turbulent and Perfect gas turbulent) of the Ma Number distribution and temperature distributions are shown in Figures 1 and 2 respectively.

Allgood, Daniel C.↗

The coupling of the neutron transport application RATTLESNAKE to the nuclear fuels performance application BISON under the MOOSE framework

The MOOSE neutron transport application RATTLESNAKE was coupled to the fuels performance application BISON to provide a higher fidelity tool for fuel performance simulation. This project is motivated by the desire to couple a high fidelity core analysis program (based on the self-adjoint angular flux equations) to a high fidelity fuel performance program, both of which can simulate on unstructured meshes. RATTLESNAKE solves self-adjoint angular flux transport equation and provides a sub-pin level resolution of the multigroup neutron flux with resonance treatment during burnup or a fast transient. BISON solves the coupled thermomechanical equations for the fuel on a sub-millimeter scale. Both applications are able to solve their respective systems on aligned and unaligned unstructured finite element meshes. The power density and local burnup was transferred from RATTLESNAKE to BISON with the MOOSE Multiapp transfer system. Multiple depletion cases were run with one-way data transfer from RATTLESNAKE to BISON. The eigenvalues are shown to agree well with values obtained from the lattice physics code DRAGON. The one-way data transfer of power density is shown to agree with the power density obtained from an internal Lassman-style model in BISON.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Akuna - Integrated Toolsets Supporting Advanced Subsurface Flow and Transport Simulations for Environmental Management

A next generation open source subsurface simulator and user environment for environmental management is being developed through a collaborative effort across Department of Energy National Laboratories. The flow and transport simulator, Amanzi, will be capable of modeling complex subsurface environments and processes using both unstructured and adaptive meshes at very fine spatial resolutions that require supercomputing-scale resources. The user environment, Akuna, provides users with a range of tools to manage environmental and simulator data sets, create models, manage and share simulation data, and visualize results. Underlying the user interface are core toolsets that provide algorithms for sensitivity analysis, parameter estimation, and uncertainty quantification. Akuna is open-source, cross platform software that is initially being demonstrated on the Hanford BC Cribs remediation site. In this paper, we describe the emerging capabilities of Akuna and illustrate how these are being applied to the BC Cribs site.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

AKUNA – INTEGRATED TOOLSETS SUPPORTING ADVANCED SUBSURFACE FLOW AND TRANSPORT SIMULATIONS FOR ENVIRONMETAL MANAGEMENT

A next generation open source subsurface simulator and user environment for environmental management is being developed through a collaborative effort across Department of Energy National Laboratories. The flow and transport simulator, Amanzi, will be capable of modeling complex subsurface environments and processes using both unstructured and adaptive meshes at very fine spatial resolutions that require supercomputing-scale resources. The user environment, Akuna, provides users with a range of tools to manage environmental and simulator data sets, create models, manage and share simulation data, and visualize results. Underlying the user interface are core toolsets that provide algorithms for sensitivity analysis, parameter estimation, and uncertainty quantification. Akuna is open-source, cross platform software that is initially being demonstrated on the Hanford BC Cribs remediation site. In this paper, we describe the emerging capabilities of Akuna and illustrate how these are being applied to the BC Cribs site.

user interface, data management, uncertainty quant↗

Parallel Grid Manipulations in Earth Science Calculations

The National Aeronautics and Space Administration (NASA) Data Assimilation Office (DAO) at the Goddard Space Flight Center is moving its data assimilation system to massively parallel computing platforms. This parallel implementation of GEOS DAS will be used in the DAO's normal activities, which include reanalysis of data, and operational support for flight missions. Key components of GEOS DAS, including the gridpoint-based general circulation model and a data analysis system, are currently being parallelized. The parallelization of GEOS DAS is also one of the HPCC Grand Challenge Projects. The GEOS-DAS software employs several distinct grids. Some examples are: an observation grid- an unstructured grid of points at which observed or measured physical quantities from instruments or satellites are associated- a highly-structured latitude-longitude grid of points spanning the earth at given latitude-longitude coordinates at which prognostic quantities are determined, and a computational lat-lon grid in which the pole has been moved to a different location to avoid computational instabilities. Each of these grids has a different structure and number of constituent points. In spite of that, there are numerous interactions between the grids, e.g., values on one grid must be interpolated to another, or, in other cases, grids need to be redistributed on the underlying parallel platform. The DAO has designed a parallel integrated library for grid manipulations (PILGRIM) to support the needed grid interactions with maximum efficiency. It offers a flexible interface to generate new grids, define transformations between grids and apply them. Basic communication is currently MPI, however the interfaces defined here could conceivably be implemented with other message-passing libraries, e.g., Cray SHMEM, or with shared-memory constructs. The library is written in Fortran 90. First performance results indicate that even difficult problems, such as above-mentioned pole rotation- a sparse interpolation with little data locality between the physical lat-lon grid and a pole rotated computational grid- can be solved efficiently and at the GFlop/s rates needed to solve tomorrow's high resolution earth science models. In the subsequent presentation we will discuss the design and implementation of PILGRIM as well as a number of the problems it is required to solve. Some conclusions will be drawn about the potential performance of the overall earth science models on the supercomputer platforms foreseen for these problems.

Sawyer, W.↗

MPAS-Analysis

MPAS-Analysis is a software package designed to process and analyze output data of the Model for Prediction Across Scales (MPAS) Ocean, Sea-Ice and Land-Ice model components. It is built primarily on the python language, with remapping of model unstructured-grid fields onto regular grids performed through the ESMF package, and with the computation of climatologies performed through the NCO package. The main purpose of MPAS-Analysis is to analyze the MPAS simulations in terms of ocean and ice metrics, in comparison with available observations, and therefore provide an accurate model evaluation and simulation assessment.

Veneziani, Milena↗

Characterizing the Non-linear Interactions Between Tide, Storm Surge, and River Flow in the Delaware Bay Estuary, United States

Low-lying coastal areas in the mid-Atlantic region are prone to compound flooding resulting from the co-occurrence of river floods and coastal storm surges. To better understand the contribution of non-linear tide-surge-river interactions to compound flooding, the unstructured-grid Finite Volume Community Ocean Model was applied to simulate coastal storm surge and flooding in the Delaware Bay Estuary in the United States. The model was validated with tide gauge data in the estuary for selected hurricane events. Non-linear interactions between tide-surge-river were investigated using a non-stationary tidal analysis method, which decomposes the interactions’ components at the frequency domain. Model results indicated that tide-river interactions damped semidiurnal tides, while the tide-surge interactions mainly influenced diurnal tides. Tide-river interactions suppressed the water level upstream while tide-surge interaction increased the water level downstream, which resulted in a transition zone of damping and enhancing effects where the tide-surge-river interaction was prominent. Evident compound flooding was observed as a result of non-linear tide-surge-river interactions. Furthermore, sensitivity analysis was carried out to evaluate the effect of river flooding on the non-linear interactions. The transition zone of damping and enhancing effects shifted downstream as the river flow rate increased.

54 ENVIRONMENTAL SCIENCES↗

Development, Verification and Validation of Parallel, Scalable Volume of Fluid CFD Program for Propulsion Applications

There are many instances involving liquid/gas interfaces and their dynamics in the design of liquid engine powered rockets such as the Space Launch System (SLS). Some examples of these applications are: Propellant tank draining and slosh, subcritical condition injector analysis for gas generators, preburners and thrust chambers, water deluge mitigation for launch induced environments and even solid rocket motor liquid slag dynamics. Commercially available CFD programs simulating gas/liquid interfaces using the Volume of Fluid approach are currently limited in their parallel scalability. In 2010 for instance, an internal NASA/MSFC review of three commercial tools revealed that parallel scalability was seriously compromised at 8 cpus and no additional speedup was possible after 32 cpus. Other non-interface CFD applications at the time were demonstrating useful parallel scalability up to 4,096 processors or more. Based on this review, NASA/MSFC initiated an effort to implement a Volume of Fluid implementation within the unstructured mesh, pressure-based algorithm CFD program, Loci-STREAM. After verification was achieved by comparing results to the commercial CFD program CFD-Ace+, and validation by direct comparison with data, Loci-STREAM-VoF is now the production CFD tool for propellant slosh force and slosh damping rate simulations at NASA/MSFC. On these applications, good parallel scalability has been demonstrated for problems sizes of tens of millions of cells and thousands of cpu cores. Ongoing efforts are focused on the application of Loci-STREAM-VoF to predict the transient flow patterns of water on the SLS Mobile Launch Platform in order to support the phasing of water for launch environment mitigation so that vehicle determinantal effects are not realized.

West, Jeff↗

Multi-region unstructured volume segmentation using tetrahedron filling

Segmentation is one of the most common operations in image processing, and while there are several solutions already present in the literature, they each have their own benefits and drawbacks that make them well-suited for some types of data and not for others. We focus on the problem of breaking an image into multiple regions in a single segmentation pass, while supporting both voxel and scattered point data. To solve this problem, we begin with a set of potential boundary points and use a Delaunay triangulation to complete the boundaries. We use heuristic- and interaction-driven Voronoi clustering to find reasonable groupings of tetrahedra. Apart from the computation of the Delaunay triangulation, our algorithm has linear time complexity with respect to the number of tetrahedra.

97 MATHEMATICS AND COMPUTING↗

Demonstration of acoustic monitoring for structural health of microreactors: Through use of neural networks and resonant ultrasound spectroscopy

Nuclear microreactors prioritize modularity and portability and are intended to be a cost-effective technology for non-conventional nuclear markets. As such, the development of microreactors into a safe and feasible solution for energy security applications will necessitate the development of non-destructive technologies to monitor the integrity of inaccessible reactors components during operation. This demonstration applies linear and nonlinear acoustic techniques, in combination with machine learning, to detect and classify mechanical changes (stress and damage) in a test article which are broadly representative of potential operating challenges within a functioning microreactor. All necessary data has been collected for this demonstration, with minor experimental issues identified that can be addressed in follow-on work. Motivated by the expected conditions within a functioning microreactor, we have demonstrated our monitoring techniques on a core-block-like test article using an unstructured excitation source that approximates the noisy acoustic environment expected during reactor operation. At all stress states mechanically applied to the test article, a machine learning model using an artificial neural network was able to classify with 100% accuracy whether a 3D laser vibrometry point measurement was made on an intact or artificially defective test article. Further, model predictions about whether the defect interface was rough or smooth were 95% accurate, indicating the ability of acoustic techniques to recover defect characteristics. Resonant ultrasound spectroscopy (RUS) was also applied to the dataset to provide further quantitative insights about material properties. RUS analysis was ultimately hampered by several minor experimental and data issues, limiting results to certain cases for this demonstration. Last, analysis using nonlinear RUS exhibited sensitivity to changing levels of applied stresses for each intact and defective state. As presented in this demonstration, acoustic monitoring exhibits sensitivity to stress changes, which are of concern due to high thermal gradients expected during startup and operation. Further, our techniques distinguish between measurements made on intact and damaged test articles.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Explicit quantification of residence and flushing times in the Salish Sea using a sub-basin scale shoreline resolving model

The Salish Sea, located in the Pacific Northwest region of North America has complex currents and circulation features distributed over numerous interconnected deep basins with islands. Increased risk of exposure to oil spills and untreated wastewater from maritime emergencies and treatment plant failures have led to a need for quantifying residence and flushing characteristics at a sub-basin scale using a shoreline resolving hydrodynamic model. An unstructured grid model of the Salish Sea was developed using FVCOM with a ≈75-100m shoreline resolution. In addition to 7 tides and 23 salinity and temperature monitoring stations, an extensive currents data set from 135 stations collected over a span of three years was used for skill assessment and validation. Explicit forward computations were then conducted to define and quantify residence and flushing times in various sub-basins of interest using (a) Lagrangian particles and (b) Numerical/virtual dye experiments. The results in most basins show expected seasonal variability with longer flushing time associated with summer lower tides and lower freshwater inflows. However, contrary to expectation, flushing time is significantly longer in wintertime in large fjord-like basins such as Hood Canal (≈138 days), likely due to increased stratification and reduced mixing. The flushing time for the Puget Sound region of the Salish Sea is ≈115 days, while Georgia Basin is 240 days when analyzed as stand-alone basins with zero background concentrations. When examined as part of the flushing of the entire Puget Sound filled with virtual dye, the compounded flushing times for embedded sub-basins can be significantly longer in order of magnitudes and largely dictated by the flushing time of Puget Sound. The computed residence and flushing time scales tabulated over 36 sub-basins provide an improved understanding of water renewal in the system, informing pollution management actions.

54 ENVIRONMENTAL SCIENCES↗

The temperature structure, mass, and energy flow in the corona and inner solar wind

Remote-sensing and in situ data are used to constrain a radiative energy balance model in order to study the radial variations of coronal temperatures, densities, and outflow speeds in several types of coronal holes and in an unstructured quiet region of the corona. A one-fluid solar wind model is used which takes into account the effects of radiative and inward conductive losses in the low corona and the chromospheric-coronal transition region. The results show that the total nonradiative energy input in magnetically open coronal regions is 5 + or - 10 to the 5th ergs/sq cm, and that most of the energy heating the coronal plasma is dissipated within 2 solar radii of the solar surface.

Withbroe, George L.↗

Unstructured grids on SIMD torus machines

Unstructured grids lead to unstructured communication on distributed memory parallel computers, a problem that has been considered difficult. Here, we consider adaptive, offline communication routing for a SIMD processor grid. Our approach is empirical. We use large data sets drawn from supercomputing applications instead of an analytic model of communication load. The chief contribution of this paper is an experimental demonstration of the effectiveness of certain routing heuristics. Our routing algorithm is adaptive, nonminimal, and is generally designed to exploit locality. We have a parallel implementation of the router, and we report on its performance.

Bjorstad, Petter E.↗

An Extensible Schema-less Database Framework for Managing High-throughput Semi-Structured Documents

Object-Relational database management system is an integrated hybrid cooperative approach to combine the best practices of both the relational model utilizing SQL queries and the object oriented, semantic paradigm for supporting complex data creation. In this paper, a highly scalable, information on demand database framework, called NETMARK is introduced. NETMARK takes advantages of the Oracle 8i object-relational database using physical addresses data types for very efficient keyword searches of records for both context and content. NETMARK was originally developed in early 2000 as a research and development prototype to solve the vast amounts of unstructured and semi-structured documents existing within NASA enterprises. Today, NETMARK is a flexible, high throughput open database framework for managing, storing, and searching unstructured or semi structured arbitrary hierarchal models, XML and HTML.

Maluf, David A.↗

Roofline Analysis in the Intel® Advisor to Deliver Optimized Performance for applications on Intel® Xeon Phi™ Processor

In this session we show, in two case studies, how the roofline feature of Intel Advisor has been utilized to optimize the performance of kernels of the XGC1 and PICSAR codes in preparation for Intel Knights Landing architecture. The impact of the implemented optimizations and the benefits of using the automatic roofline feature of Intel Advisor to study performance of large applications will be presented. This demonstrates an effective optimization strategy that has enabled these science applications to achieve up to 4.6 times speed-up and prepare for future exascale architectures. # Goal/Relevance of Session The roofline model [1,2] is a powerful tool for analyzing the performance of applications with respect to the theoretical peak achievable on a given computer architecture. It allows one to graphically represent the performance of an application in terms of operational intensity, i.e. the ratio of flops performed and bytes moved from memory in order to guide optimization efforts. Given the scale and complexity of modern science applications, it can often be a tedious task for the user to perform the analysis on the level of functions or loops to identify where performance gains can be made. With new Intel tools, it is now possible to automate this task, as well as base the estimates of peak performance on measurements rather than vendor specifications. The goal of this session is to demonstrate how the roofline feature of Intel Advisor can be used to balance memory vs. computation related optimization efforts and effectively identify performance bottlenecks. A series of typical optimization techniques: cache blocking, structure refactoring, data alignment, and vectorization illustrated by the kernel cases will be addressed. # Description of the codes ## XGC1 The XGC1 code [3] is a magnetic fusion Particle-In-Cell code that uses an unstructured mesh for its Poisson solver that allows it to accurately resolve the edge plasma of a magnetic fusion device. After recent optimizations to its collision kernel [4], most of the computing time is spent in the electron push (pushe) kernel, where these optimization efforts have been focused. The kernel code scaled well with MPI+OpenMP but had almost no automatic compiler vectorization, in part due to indirect memory addresses and in part due to low trip counts of low-level loops that would be candidates for vectorization. Particle blocking and sorting have been implemented to increase trip counts of low-level loops and improve memory locality, and OpenMP directives have been added to vectorize compute-intensive loops that were identified by Advisor. The optimizations have improved the performance of the pushe kernel 2x on Haswell processors and 1.7x on KNL. The KNL node-for-node performance has been brought to within 30% of a NERSC Cori phase I Haswell node and we expect to bridge this gap by reducing the memory footprint of compute intensive routines to improve cache reuse. ## PICSAR is a Fortran/Python high-performance Particle-In-Cell library targeting at MIC architectures first designed to be coupled with the PIC code WARP for the simulation of laser-matter interaction and particle accelerators. PICSAR also contains a FORTRAN stand-alone kernel for performance studies and benchmarks. A MPI domain decomposition is used between NUMA domains and a tile decomposition (cache-blocking) handled by OpenMP has been added for shared-memory parallelism and better cache management. The so-called current deposition and field gathering steps that compose the PIC time loop constitute major hotspots that have been rewritten to enable more efficient vectorization. Particle communications between tiles and MPI domain has been merged and parallelized. All considered, these improvements provide speedups of 3.1 for order 1 and 4.6 for order 3 interpolation shape factors on KNL configured in SNC4 quadrant flat mode. Performance is similar between a node of cori phase 1 and KNL at order 1 and better on KNL by a factor 1.6 at order 3 with the considered test case (homogeneous thermal plasma).

97 MATHEMATICS AND COMPUTING↗