Results for “gpu” — Page 39

NASA NTRSDate not supplied

Performance and Portability of a Linear Solver Across Emerging Architectures

A linear solver algorithm used by a large-scale unstructured-grid computational fluid dynamics application is examined for a broad range of familiar and emerging architectures. Efficient implementation of a linear solver is challenging on recent CPUs offering vector architectures. Vector loads and stores are essential to effectively utilize available memory bandwidth on CPUs, and maintaining performance across different CPUs can be difficult in the face of varying vector lengths offered by each. A similar challenge occurs on GPU architectures, where it is essential to have coalesced memory accesses to utilize memory bandwidth effectively. In this work, we demonstrate that restructuring a computation, and possibly data layout, with regard to architecture is essential to achieve optimal performance by establishing a performance benchmark for each target architecture in a low level language such as vector intrinsics or CUDA. In doing so, we demonstrate how a linear solver kernel can be mapped to Intel® Xeon™ and Xeon Phi™, Marvell® ThunderX2®, NEC® SX-Aurora™ TSUBASA Vector Engine, and NVIDIA® and AMD® GPUs. We further demonstrate that the required code restructuring can be achieved in higher level programming environments such as OpenACC, OCCA, and Intel® OneAPI™/SYCL, and that each generally results in optimal performance on the target architecture. Relative performance metrics for all implementations are shown, and subjective ratings for ease of implementation and optimization are suggested.

Programming models↗

NASA NTRSDate not supplied

Implicit Thermochemical Nonequilibrium Compressible Flow Simulations on Unstructured Grids Using GPUs

As next-generation exascale-class systems arrive, existing software must be updated accordingly to effectively utilize these systems. For high concurrency and energy efficiency, many of these systems utilize GPU architectures. In this work, we present a CUDA C++ implementation of FUN3D's thermochemical nonequilibrium capability for turbulent flows. Efficiency is demonstrated at scale using the Summit system at the Oak Ridge Leadership Computing Facility which is representative of future exascale systems. This work enables faster, higher fidelity, and scale-resolving simulations of thermochemical nonequilibrium flows including reentry, hypersonics, and combustion.

CFD, GPU, HPC, Hypersonics, Chemistry↗

NASA NTRSDate not supplied

Implicit Thermochemical Nonequilibrium Flow Simulations on Unstructured Grids using GPUs

Thermochemical nonequilibrium flow simulation capabilities have been previously implemented, verified, and validated for central processing unit (CPU) systems in NASA’s unstructured-grid computational fluid dynamics solver FUN3D. Many exascale-class high-performance computing systems will rely on graphics processing unit (GPU) architectures for high throughput and energy efficiency; thus, CPU-based scientific computing software unable to effectively utilize these systems must be updated. In this work, we present a CUDA C++ implementation of FUN3D’s thermochemical nonequilibrium flow simulation capabilities targeting NVIDIA Tesla GPUs. An overview of the porting and optimization strategy is described and performance comparisons with other recent architectures are presented. Scaling to thousands of GPUs is demonstrated, yielding computational performance equivalent to that of several million CPU cores. The implementation enables efficient, high-fidelity, scale-resolving simulations of thermochemical nonequilibrium flows for many applications including atmospheric entry, hypersonics, and combustion.

GPU↗

NASA NTRSDate not supplied

Computational Investigation of the Effect of Chemistry on Mars Supersonic Retropropulsion Environments

Retropropulsion ground tests require significant compromises on physical scale, instrumentation, configuration, and environments. Matching the full Martian environment is simply not possible on Earth. Ground tests of retropropulsion configurations thus far have neglected effects of chemistry due to physical constraints of wind tunnel models and facilities; most experiments use inert simulant gases at relatively low temperatures. As such, a strong reliance on high-fidelity computational analyses is required to expand the knowledge of retropropulsion aerodynamics. In this work, we investigate the effects of chemistry using scale-resolving computational fluid dynamics (CFD) with finite-rate chemistry and a graphics processing unit (GPU)-enabled implementation of the NASA FUN3D flow solver on a human-scale Mars lander concept at supersonic freestream conditions. Results are compared to previous perfect gas simulations.

retropropulsion↗

NASA NTRSDate not supplied

Computational Investigation of the Effect of Chemistry on Mars Supersonic Retropropulsion Environments

Retropropulsion ground tests require significant compromises on physical scale, instrumentation, configuration, and environments. Matching the full Martian environment is simply not possible on Earth. Ground tests of retropropulsion configurations thus far have neglected effects of chemistry due to physical constraints of wind tunnel models and facilities; most experiments use inert simulant gases at relatively low temperatures. As such, a strong reliance on high-fidelity computational analyses is required to expand the knowledge of retropropulsion aerodynamics. In this work, we investigate the effects of chemistry using scale-resolving computational fluid dynamics (CFD) with finite-rate chemistry and a graphics processing unit (GPU)-enabled implementation of the NASA FUN3D flow solver on a human-scale Mars lander concept at supersonic freestream conditions. Results are compared to previous perfect gas simulations.

retropropulsion↗

NASA NTRSDate not supplied

OVERFLOW on GPUs: Progress and Lessons Learned

Overview - Goals of OVERFLOW GPU Port - Miniapps - Lessons Learned - Current Status - What is Coming

GPU↗

NASA NTRSDate not supplied

A Multi-Architecture Approach for Implicit Computational Fluid Dynamics on Unstructured Grids

High-performance computing (HPC) architectures are trending toward manycore paradigms such as graphics processing units (GPUs). Approximately half of the top 100 publicly disclosed supercomputers in the world utilize GPU accelerators for performance. This is in contrast to a decade ago, where there were only a few such machines in the top 100. It is not currently possible to compile and run legacy central processing unit (CPU) software efficiently on GPUs without significant refactoring. Though a number of frameworks offering performance portability exist, none offer a standardized specification that is supported by all major hardware vendors. Additionally, experiences show that obtaining a high percentage of peak performance often requires architecture-specific code. This work details a pragmatic multi-architecture computational fluid dynamics library focused on aerospace problems across the speed range from low subsonic to hypersonic flows involving thermochemical nonequilibrium. A thin abstraction layer above NVIDIA CUDA C++ is utilized, which enables primarily single-source software currently capable of running efficiently on multicore CPUs, NVIDIA GPUs, AMD GPUs, and Intel GPUs. Results on various problems of interest across the speed range are presented and performance is compared between various architectures.

GPU↗

NASA NTRSDate not supplied

A Multi-Architecture Approach for Implicit Computational Fluid Dynamics on Unstructured Grids

High-performance computing (HPC) architectures are trending toward manycore paradigms such as graphics processing units (GPUs). Approximately half of the top 100 publicly disclosed supercomputers in the world utilize GPU accelerators for performance. This is in contrast to a decade ago, where there were only a few such machines in the top 100. It is not currently possible to compile and run legacy central processing unit (CPU) software efficiently on GPUs without significant refactoring. Though a number of frameworks offering performance portability exist, none offer a standardized specification that is supported by all major hardware vendors. Additionally, experiences show that obtaining a high percentage of peak performance often requires architecture-specific code. This work details a pragmatic multi-architecture computational fluid dynamics library focused on aerospace problems across the speed range from low subsonic to hypersonic flows involving thermochemical nonequilibrium. A thin abstraction layer above NVIDIA CUDA C++ is utilized, which enables primarily single-source software currently capable of running efficiently on multicore CPUs, NVIDIA GPUs, AMD GPUs, and Intel GPUs. Results on various problems of interest across the speed range are presented and performance is compared between various architectures.

GPU↗

NASA NTRS2023

Dynamically Rendering Rough Terrain with Minimal Memory Overhead

Rendering highly detailed terrain is a process with the potential to consume a great deal of a computer’s random access memory (RAM). In a browser-based application, this resource is limited even further, leading to the necessity to use alternative methods of rendering the large amount of data needed for high detail. This report describes one such method that places the onus of rendering on the speed of the graphics processing unit (GPU) rather than on the computer’s memory. By removing attribute buffers, which contribute greatly to memory costs, from the rendering pipeline and generating the requisite attributes on the fly using a heightmap texture instead, it is estimated that memory usage can be cut down to one-sixth that of the previous method.

Visualization↗

DOE OSTI2022

SimNet: Accurate and High-Performance Computer Architecture Simulation using Deep Learning

While cycle-accurate simulators are essential tools for architecture research, design, and development, their practicality is limited by an extremely long time-to-solution for realistic applications under investigation. This work describes a concerted effort, where machine learning (ML) is used to accelerate microarchitecture simulation. First, an ML-based instruction latency prediction framework that accounts for both static instruction properties and dynamic processor states is constructed. Then, a GPU-accelerated parallel simulator is implemented based on the proposed instruction latency predictor, and its simulation accuracy and throughput are validated and evaluated against a state-of-the-art simulator. Leveraging modern GPUs, the ML-based simulator outperforms traditional CPU-based simulators significantly.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2021

Direct Chemical Vapor Deposition Synthesis of Porous Single-Layer Graphene Membranes with High Gas Permeances and Selectivities

Single-layer graphene containing molecular-sized in-plane pores is regarded as a promising membrane material for high-performance gas separations due to its atomic thickness and low gas transport resistance. However, typical etching-based pore generation methods cannot decouple pore nucleation and pore growth, resulting in a trade-off between high areal pore density and high selectivity. In contrast, intrinsic pores in graphene formed during chemical vapor deposition are not created by etching. Therefore, intrinsically porous graphene can exhibit high pore density while maintaining its gas selectivity. Here in this work, the density of intrinsic graphene pores is systematically controlled for the first time, while appropriate pore sizes for gas sieving are precisely maintained. As a result, single-layer graphene membranes with the highest H 2 /CH 4 separation performances recorded to date (H 2 permeance > 4000 GPU and H 2 /CH 4 selectivity > 2000) are fabricated by manipulating growth temperature, precursor concentration, and non-covalent decoration of the graphene surface. Moreover, it is identified that nanoscale molecular fouling of the graphene surface during gas separation where graphene pores are partially blocked by hydrocarbon contaminants under experimental conditions, controls both selectivity and temperature dependent permeance. Overall, the direct synthesis of porous single-layer graphene exploits its tremendous potential as high-performance gas-sieving membranes.

chemical vapor deposition↗

DOE OSTI2021

Design and operational considerations of catalytic membrane reactors for ammonia synthesis

Production of ammonia using hydrogen derived from renewable electricity instead of hydrocarbon reforming would dramatically reduce the carbon footprint of this commodity chemical. Novel technologies such as catalytic membrane reactors may potentially be more compatible with distributed ammonia production than the conventional Haber-Bosch process. Here, a reactor model is developed based on integrating a standard industrial iron catalyst into a catalytic membrane reactor (CMR) equipped with an inorganic membrane that is selective to NH 3 over N 2 /H 2 . CMR performance is studied as functions of wide ranges of membrane properties and operating conditions. Conversion and ammonia recovery are dictated principally by the ammonia permeance, and the benefits by using membranes become significant above 100 GPU = 3.4 × 10 –8 mol m –2 s –1 Pa –1 . To be effective, the CMR requires a minimum selectivity for ammonia of 10 over both nitrogen and hydrogen, and purity scales with the effective selectivity. Increasing the pressure of operation significantly improves all metrics, and at P = 30 bar with a quality membrane ammonia is almost completely recovered, enabling direct recycle of un-reacted hydrogen and nitrogen without need for recompression. Temperature drives conversion and scales monotonically without thermodynamic limitations in a CMR. Alternatively, the temperature may be reduced as low as 300°C while achieving conversion levels surpassing equilibrium limits at T = 400°C in a conventional reactor.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2023

Appealing sheath‐core spun high‐performance composite carbon molecular sieve membranes

Abstract Carbon molecular sieve (CMS) membranes are attractive candidates to meet requirements for challenging gas separations. The added ability to maintain such intrinsic properties in an asymmetric morphology with a structure that we term a “Pseudo Wheel+Hub & Spoke” asymmetric form offers new opportunities. For CMS membrane, specifically, the structure provides both selective layer support and low flow resistance even for high feed pressures and fluxes in CO 2 removal from natural gas. This capability is unavailable to even rigid glassy polymers due to the much higher modulus of CMS materials. Combining precursor asymmetric hollow fiber formation and optimized pyrolysis creates a defect free CMS proof‐of‐concept membrane for this application. Facile formation of the sheath‐core spun precursor with a 6FDA‐DAM sheath and Matrimid® core also avoids the need to seal defects before or after the carbonization of the precursors. The composite CMS membrane shows CO 2 /CH 4 (50 : 50) mixed gas feed with an attractive CO 2 /CH 4 selectivity of 64.3 and CO 2 permeance of 232 GPU at 35 °C. A key additional benefit of the approach is reduction in use of the more costly high performance 6FDA‐DAM in a composite sheath‐core CMS membrane with the “Pseudo Wheel+Hub & Spoke” structure.

Cao, Yuhe↗

DOE OSTI2025

Synergistic Tuning of Microstructure and Morphology in Carbon Molecular Sieve Hollow Fibers for Propylene/Propane Separation

Abstract Asymmetric carbon molecular sieve (CMS) hollow fiber membranes with tunable micro‐ and macro‐structural morphologies for energy efficient propylene‐propane separation are reported here. A sub‐glass transition temperature (sub‐Tg) thermal oxidative crosslinking strategy enables simultaneous optimization of the intrinsic molecular sieving properties while also reducing the thickness of the CMS “skin” derived from the 6FDA : BPDA/DAM polyimide precursors. Such synergistic tuning of CMS microstructure and macroscopic morphology of CMS hollow fibers enables significantly increased propylene permeance (reaching 186.5 GPU) while maintaining an appealing propylene/propane selectivity of 13.3 for 50/50 propylene/propane mixed gas feeds. Our findings reveal a more refined and versatile tool than available with previous O 2 ‐doping pretreatments. The advanced approach here should be broadly useful to other polyimide precursors and diverse gas pairs.

Liu, Zhongyun [School of Chemical &amp, Biomolecul↗

DOE OSTI2023

Appealing sheath-core spun high-performance composite carbon molecular sieve membranes

Carbon molecular sieve (CMS) membranes are attractive candidates to meet requirements for challenging gas separations. The added ability to maintain such intrinsic properties in an asymmetric morphology with a structure that we term a “Pseudo Wheel+Hub & Spoke” asymmetric form offers new opportunities. For CMS membrane, specifically, the structure provides both selective layer support and low flow resistance even for high feed pressures and fluxes in CO 2 removal from natural gas. Further, this capability is unavailable to even rigid glassy polymers due to the much higher modulus of CMS materials. Combining precursor asymmetric hollow fiber formation and optimized pyrolysis creates a defect free CMS proof-of-concept membrane for this application. Facile formation of the sheath-core spun precursor with a 6FDA-DAM sheath and Matrimid® core also avoids the need to seal defects before or after the carbonization of the precursors. The composite CMS membrane shows CO 2 /CH 4 (50 : 50) mixed gas feed with an attractive CO 2 /CH 4 selectivity of 64.3 and CO 2 permeance of 232 GPU at 35 °C. A key additional benefit of the approach is reduction in use of the more costly high performance 6FDA-DAM in a composite sheath-core CMS membrane with the “Pseudo Wheel+Hub & Spoke” structure.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2024

Synergistic Tuning of Microstructure and Morphology in Carbon Molecular Sieve Hollow Fibers for Propylene/Propane Separation

Abstract Asymmetric carbon molecular sieve (CMS) hollow fiber membranes with tunable micro‐ and macro‐structural morphologies for energy efficient propylene‐propane separation are reported here. A sub‐glass transition temperature (sub‐Tg) thermal oxidative crosslinking strategy enables simultaneous optimization of the intrinsic molecular sieving properties while also reducing the thickness of the CMS “skin” derived from the 6FDA : BPDA/DAM polyimide precursors. Such synergistic tuning of CMS microstructure and macroscopic morphology of CMS hollow fibers enables significantly increased propylene permeance (reaching 186.5 GPU) while maintaining an appealing propylene/propane selectivity of 13.3 for 50/50 propylene/propane mixed gas feeds. Our findings reveal a more refined and versatile tool than available with previous O 2 ‐doping pretreatments. The advanced approach here should be broadly useful to other polyimide precursors and diverse gas pairs.

Liu, Zhongyun↗

DOE OSTI2020

Experiences in porting mini-applications to OpenACC and OpenMP on heterogeneous systems

This article studies mini-applications—Minisweep, GenASiS , GPP, and FF—that use computational methods commonly encountered in HPC. We have ported these applications to develop OpenACC and OpenMP versions, and evaluated their performance on Titan (Cray XK7 with K20x GPUs), Cori (Cray XC40 with Intel KNL), Summit (IBM AC922 with Volta GPUs), and Cori-GPU (Cray CS-Storm 500NX with Intel Skylake and Volta GPUs). Our goals are for these new ports to be useful to both application and compiler developers, to document and describe the lessons learned and the methodology to create optimized OpenMP and OpenACC versions, and to provide a description of possible migration paths between the two specifications. Cases where specific directives or code patterns result in improved performance for a given architecture are highlighted. Here, we also include discussions of the functionality and maturity of the latest compilers available on the above platforms with respect to OpenACC or OpenMP implementations.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2021

FPGA-based HPC accelerators: An evaluation on performance and energy efficiency

Hardware specialization is a promising direction for the future of digital computing. Reconfigurable technologies enable hardware specialization with modest non-recurring engineering cost, but their performance and energy efficiency compared to state-of-the-art processor architectures remain an open question. In this article, we use FPGAs to evaluate the benefits of building specialized hardware for numerical kernels found in scientific applications. In order to properly evaluate performance, we not only compare Intel Arria 10 and Xilinx U280 performance against Intel Xeon, Intel Xeon Phi, and NVIDIA V100 GPUs, but we also extend the Empirical Roofline Toolkit (ERT) to FPGAs in order to assess our results in terms of the Roofline model. We show design optimization and tuning techniques for peak FPGA performance at reasonable hardware usage and power consumption. As FPGA peak performance is known to be far less than that of a GPU, we also benchmark the energy efficiency of each platform for the scientific kernels comparing against microbenchmark and technological limits. Results show that while FPGAs struggle to compete in absolute terms with GPUs on memory- and compute-intensive kernels, they require far less power and can deliver nearly the same energy efficiency.

97 MATHEMATICS AND COMPUTING↗