Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “performance portable”

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 559 records · Page 31

FY2022 Q4: Demonstrate multi-turbine simulation with hybrid-structured/unstructured-moving-grid software stack running primarily on GPUs and propose improvements for successful KPP-2 [Poster]

Milestone accomplishments staged the ExaWind team for successful completion of KPP-2 challenge problem in FY23, which requires the simulation on Frontier of at least four MW-scale turbines in an atmospheric boundary layer with at least 20B gridpoints. The ExaWind project and software stack is many faceted, with team members working on multiple areas, including linear-system solvers (Trilinos, hypre, AMReX), overset meshes, turbulence modeling, and in situ visualization, all with an aim for high fidelity predictions and performance portability. This milestone marks significant improvements on many fronts and provides the team with a pathway to exascale wind farm simulations in FY23.

17 WIND ENERGY↗

Extending PETSc's Composable Hierarchical Solvers (Final Technical Report)

This report documents research activities conducted at CU Boulder as part of Extending PETSc’s Composable Hierarchical Solvers, which has been part of a collaboration with Argonne National Laboratory (separate award). Our work has focused on performance-portable end-to-end GPU solvers demonstrated via exemplary applications in nonlinear fluid and structural mechanics. We describe advances in algorithmic composition and analysis in the context of these applications, but the implementations are fully documented and decoupled, and in use by other projects. We believe the vertical integration achieved through collaboration with ECP’s CEED and the PSAAP center at CU was necessary to take risks with data structures and algorithms.

42 ENGINEERING↗

To Interoperability And Beyond: Interoperable Types Through the Promises of C and C++ and ABI Abuse [Slides]

This presentation presents a technique that allows passing of Fortran nested-type hierarchies interoperably to C++. The technique is motivated by the Eulerian Application Project’s need to port code from Fortran to C++ to utilize the Kokkos performance portability library. The resulting method allows hierarchies of types to become interoperable while at the same time transforming Fortran array members of the original Fortran type into Kokkos::Views in the resulting C++ type.

97 MATHEMATICS AND COMPUTING↗

Milestone 49 Report: Batched Sparse LA Phase 5 Implementation

Batched sparse linear algebra operations in general, and solvers in particular, have become the major algorithmic development activity and foremost performance engineering effort in the numerical software libraries work on modern hardware with accelerators such as GPUs. Many applications, ECP and non-ECP alike, require simultaneous solutions of many small linear systems of equations that are structurally sparse in one form or another. In order to move towards high hardware utilization levels, it is important to provide these applications with appropriate interface designs to be both functionally efficient and performance portable and give full access to the appropriate batched sparse solvers running on modern hardware accelerators prevalent across DOE supercomputing sites since the inception of ECP. To this end, we present here a summary of recent advances on the interface designs in use by HPC software libraries supporting batched sparse linear algebra and the development of sparse batched kernel codes for solvers and preconditioners. We also address the potential interoperability opportunities to keep the corresponding software portable between the major hardware accelerators from AMD, Intel, and NVIDIA, while maintaining the appropriate disclosure levels conforming to the active NDA agreements. The presented interface specifications include a mix of batched band, sparse iterative, and sparse direct solvers with their accompanying functionality that is already required by the application codes or we anticipated to be needed in the near future. This report summarizes progress in Kokkos Kernels and the xSDK libraries MAGMA, Ginkgo, hypre, PETSc, and SuperLU.

97 MATHEMATICS AND COMPUTING↗

An Adaptive Particle Tracking Algorithm for Lagrangian-Eulerian Simulations of Dispersed Multiphase Flows

The Lagrangian-Eulerian method is broadly used for simulations of dispersed multiphase flows by solving the continuous phase in the Eulerian framework while treating the dispersed phase as point particles in a Lagrangian framework. The accuracy of the Lagrangian-Eulerian method largely depends on the number of computational particles tracked for the Lagrangian phase. In this study, an adaptive Lagrangian particle tracking algorithm is proposed to balance the statistical error and the computational cost by dividing and merging the particles according to the local particle statistics in the computational domain. The computational particles are considered as weighted sampling points of the particle probability density functions (PDF) which represents a statistical equivalence of the Lagrangian phase. The algorithm is implemented as a part of a C++ library for Lagrangian particles, Grit, with performance portability to multi-core or many-core CPUs and Graphic Processing Unit (GPU) architectures. The accuracy of the proposed algorithm with two different schemes are studied through a test problem with analytical solutions for both the particle number density and momentum source term. The results are used to evaluate the proposed algorithm as well as to validate the parallel implementation.

Ge, Wenjun↗

Lagrangian simulations of spray interaction with a surface using a stochastic multi-regime impingement model

Modeling of spray impingement upon the piston or cylinder surface is important for hydrocarbon emission and soot predictions in high-fidelity simulation of internal combustion engines with direct injections. Due to the complexity of the physical process of spray impingement, current computational models heavily rely on the empirical studies which involve large uncertainties. A computational model is presented for spray impingement on solid walls in general computational fluid dynamics (CFD) simulations. The model adopts the Marmanis-Thoroddsen theory for determining the number of secondary droplets and a stochastic method for calculating the velocities of the secondary droplets. The numerical methods and parallel implementation of the impingement model are presented in detail. The model is implemented in Grit, which is a recently developed CPU/GPU performance-portable Lagrangian particle library, for accelerating spray simulations. The model and its implementation are tested for examples of spray impingements on flat and smooth surfaces with different impingement angles.

Ge, Wenjun↗

CabanaPD

CabanaPD is a meshfree peridynamics application built with Cabana and Kokkos. Kokkos enables performance portability across hardware architectures and Cabana provides particle capabilities including multi-node MPI support. The main components of CabanaPD are particle initialization, neighbor list generation, force and energy computation, time integration, and multi-node particle communication. In addition, options for creating pre-cracked regions and particle boundary conditions are available. CabanaPD currently enables two common force models: prototype microelastic brittle (PMB) and linear peridynamic solid (LPS). For both of these models, versions with and without fracture as well as linearized model options are available. CabanaPD is designed to be extensible for addition of more complex force models, boundary conditions, etc.

Reeve, Sam↗

Developing Information Power Grid Based Algorithms and Software

This exploratory study initiated our effort to understand performance modeling on parallel systems. The basic goal of performance modeling is to understand and predict the performance of a computer program or set of programs on a computer system. Performance modeling has numerous applications, including evaluation of algorithms, optimization of code implementations, parallel library development, comparison of system architectures, parallel system design, and procurement of new systems. Our work lays the basis for the construction of parallel libraries that allow for the reconstruction of application codes on several distinct architectures so as to assure performance portability. Following our strategy, once the requirements of applications are well understood, one can then construct a library in a layered fashion. The top level of this library will consist of architecture-independent geometric, numerical, and symbolic algorithms that are needed by the sample of applications. These routines should be written in a language that is portable across the targeted architectures.

Dongarra, Jack↗

Comparison of Knee and Ankle Dynamometry between NASA's X1 Exoskeleton and Biodex System 4

Pre- and post-flight dynamometry is performed on International Space Station crewmembers to characterize microgravity-induced strength changes. Strength is not assessed in flight due to hardware limitations and there is poor understanding of the time course of in-flight changes. PURPOSE: To assess the reliability of a prototype dynamometer, the X1 Exoskeleton (EXO) and its agreement with a Biodex System 4 (BIO). METHODS: Eight subjects (4 M/4 F) completed 2 counterbalanced testing sessions of knee extension/flexion (KE/KF), 1 with BIO and 1 with EXO, with repeated measures within each session in normal gravity. Test-retest reliability (test 1 and 2) and device agreement (BIO vs. EXO) were evaluated. Later, to assess device agreement for ankle plantarflexion (PF), 10 subjects (4 M/6 F) completed 3 test conditions (BIO, EXO, and BIOEXO); BIOEXO was a hybrid condition comprised of the Biodex dynamometer motor and the X1 footplate and ankle frame. Ankle comparisons were: BIO vs. BIOEXO (footplate differences), BIOEXO vs. EXO (motor differences), and BIO vs. EXO (all differences). Reliability for KE/KF was determined by intraclass correlation (ICC). Device agreement was assessed with: 1) repeated measures ANOVA, 2) a measure of concordance (rho), and 3) average difference. RESULTS: ICCs for KE/KF were 0.99 for BIO and 0.96 to 0.99 for EXO. Agreement was high for KE (concordance: 0.86 to 0.95; average differences: -7 to +9 Nm) and low to moderate for KF (concordance: 0.64 to 0.78; average differences: -4 to -29 Nm, P<0.05). BIO vs. BIOEXO PF concordance ranged from 0.89 to 0.92 and mean differences ranged from -9 to +3 Nm (BIO < BIOEXO). BIOEXO vs. EXO PF concordance ranged from 0.73 to 0.80 while mean differences were -18 to -36 Nm (BIOEXO < EXO, P<0.05). PF concordance for BIO vs. EXO was slightly lower (0.61 to 0.84) and mean differences were greater (-27 to -33 Nm; BIO < EXO, P<0.05). CONCLUSION: BIO and EXO were similarly reliable for KE and KF. KE measures produced high agreement between devices; KF did not. For ankle PF, torque differences due to the two footplates were small. However, the X1 motor reports greater torques than the Biodex motor during PF. This first prototype provides proof of concept for a reliable, robotic-based exoskeleton to perform portable dynamometry for large muscle groups of the lower body.

English, K. L.↗

Operating System Support for Fine-Grained Task Migration

With the end of clock scaling and the limited power budget available (20-30MW), future supercomputers will meet exascale performance primarily through a higher level of parallelism. Current operating (OS) and runtime (RT) systems are designed for the classical SMP model and based on the static and coarse-grained process/thread paradigm. They do not provide the required level of flexibility, especially within a single compute node, to meet the requirements imposed by exascale systems in terms of power/energy efficiency, resilience, managing concurrency and performance portability. The high level of concurrency poses new challenges specific to exascale systems that need to be addressed by novel solutions. In particular, requiring the user to manage billions of concurrent threads could easily result in poor data locality, clogged interconnection networks, unchecked propagation of soft errors, and lack of control over power/energy consumption. Equally important, managing such level of concurrency interferes with the user’s focus on the application and the algorithm. We envision that the computation will be encapsulated into fine-grained tasks that can be isolated and protected from the other tasks running in the system. Whenever a task needs to work on some data that is not stored on the local node, the OS/RT allows the task to be migrated to the node that owns the data. Each task is associated with a contained state (set or architectural registers, stack frame, running node, etc.) that describes the progress of the task and that should be moved together with the task’s code. We believe system support for task migration is a fundamental function that can ease the job of tackling several of the exascale challenges.

Gioiosa, Roberto↗

CoMD Implementation Suite in Emerging Programming Models

CoMD-Em is a software implementation suite of the CoMD [4] proxy app using different emerging programming models. It is intended to analyze the features and capabilities of novel programming models that could help ensure code and performance portability and scalability across heterogeneous platforms while improving programmer productivity. Another goal is to provide the authors and venders with some meaningful feedback regarding the capabilities and limitations of their models. The actual application is a classical molecular dynamics (MD) simulation using either the Lennard-Jones method (LJ) or the embedded atom method (EAM) for primary particle interaction. The code can be extended to support alternate interaction models. The code is expected ro run on a wide class of heterogeneous hardware configurations like shard/distributed/hybrid memory, GPU's and any other platform supported by the underlying programming model.

Haque, Riyaz↗

Kokkos Array

The Kokkos Array library implements shared-memory array data structures and parallel task dispatch interfaces for data-parallel computational kernels that are performance-portable to multicore-CPU and manycore-accelerator (e.g., GPGPU) devices.

Edwards Daniel Sunderland, Harold Carter↗

Apollo

Modern performance portability frameworks provide application developers with a flexible way to determine how to run application kernels, however, they provide no guidance as to the best configuration for a given kernel. Apollo provides a model-generation framework that, when integrated with the RAJA library, uses lightweight decision tree models to select the fastest execution configuration on a per-kernel basis

Beckingsal, David↗

A Generalized Framework for Auto-tuning Stencil Computations

This work introduces a generalized framework for automatically tuning stencil computations to achieve superior performance on a broad range of multicore architectures. Stencil (nearest-neighbor) based kernels constitute the core of many important scientific applications involving block-structured grids. Auto-tuning systems search over optimization strategies to find the combination of tunable parameters that maximizes computational efficiency for a given algorithmic kernel. Although the auto-tuning strategy has been successfully applied to libraries, generalized stencil kernels are not amenable to packaging as libraries. Studied kernels in this work include both memory-bound kernels as well as a computation-bound bilateral filtering kernel. We introduce a generalized stencil auto-tuning framework that takes a straightforward Fortran expression of a stencil kernel and automatically generates tuned implementations of the kernel in C or Fortran to achieve performance portability across diverse computer architectures.

97 MATHEMATICS AND COMPUTING↗

DPP-PMRF: Rethinking Optimization for a Probabilistic Graphical Model Using Data-Parallel Primitives

We present a new parallel algorithm for probabilistic graphical model optimization. The algorithm relies on data-parallel primitives (DPPs), which provide portable performance over hardware architecture. We evaluate results on CPUs and GPUs for an image segmentation problem. Compared to a serial baseline, we observe runtime speedups of up to 13X (CPU) and 44X (GPU). We also compare our performance to a reference, OpenMP-based algorithm, and find speedups of up to 7X (CPU).

97 MATHEMATICS AND COMPUTING↗

Towards Generic Parallel Programming in Computer Science Education with Kokkos

Parallel patterns, views, and spaces are promising abstractions to capture the programmer's intent as well as the contextual information that can be used by an underlying runtime to efficiently map software to parallel hardware. These abstractions can be valuable in cases where an algorithm must accommodate requirements of code and performance portability across hardware architectures and vendor programming models. Kokkos is a parallel programming model for host- and accelerator architectures that relies on these abstractions and targets these requirements. It consists of a pure C++ interface, a specification, and a programming library. The programming library exposes patterns and types and maps them to an underlying abstract machine model. The abstract machine model offers a generic view of parallel hardware. While Kokkos is gaining popularity in large-scale HPC applications at some DOE laboratories, we believe that the implemented concepts are of interest to a broader audience including academia as they may contribute to a generic, vendor, and architecture-independent education of parallel programming. In this work, we give an insight into the design considerations of this programming model and list important abstractions. Further, we document best practices obtained from giving virtual classes on Kokkos and give pointers to resources that the reader may consider valuable for a lecture on generic parallel programming for students with preexisting knowledge on this matter.

Ciesko, Jan↗

autoGEMM: Pushing the Limits of Irregular Matrix Multiplication on Arm Architectures

This paper presents an open-source library that pushes the limits of performance portability for irregular General Matrix Multiplication (GEMM) on the widely-used Arm architectures. Our library, autoGEMM, is designed to support a wide range of Arm processors: from edge devices to HPC-grade CPUs. autoGEMM generates optimized kernels for various hardware configurations by auto-combining fragments of autogenerated micro-kernels that employ hand-written optimizations to maximize computational efficiency. We optimize the kernel pipeline by tuning the register reuse and the data load/store overlapping. In addition, we use a dynamic tiling scheme to generate balanced tile shapes. Finally, we position autoGEMM on top of the TVM framework where our dynamic tiling scheme prunes the search space for TVM to identify the optimal combination of parameters for code optimization. Evaluations on five different classes of Arm chips demonstrate the advantages of autoGEMM. For small matrices, autoGEMM achieves 98% of peak and up to 2.0x speedup over state-of-the-art libraries such as LIBXSMM and LibShalom. For irregular matrices (i.e. tall skinny and long rectangles), autoGEMM is 1.3-2.0x faster than widely-used libraries such as OpenBLAS and Eigen. autoGEMM is publicly available at: https://github.com/wudu98/autoGEMM.

Wu, Du↗

QMCPACK v3.16.0

QMCPACK is an open-source production level many-body ab initio Quantum Monte Carlo code for computing the electronic structure of atoms, molecules, and solids with full performance portable GPU support.

Kent, Paul R. C. [Oak Ridge National Laboratory] (↗