Search NASA⌕ Search

SEARCH · Search NASA

Results for “Kernel Designs”

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 199 records · Page 11
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Initial Observations from AGR 5/6/7 Capsule 1

The fourth and final irradiation experiment in the Advanced Reactor Technologies (ART) Advanced Gas Reactor (AGR) fuel development and qualification program is designated as AGR-5/6/7. Data collected from the fabrication, irradiation, and post-irradiation examination (PIE) of this tristructural isotropic (TRISO) fuel are intended to serve as the primary data set for the qualification of this fuel for use in high-temperature gas-cooled reactors (INL 2021, Collin 2018b). However, data collected from the three preceding irradiations (i.e., AGR-1, AGR-2, and AGR-3/4) may also be used to supplement data collected from AGR-5/6/7. All components of the AGR-5/6/7 fuel (i.e., UCO kernels, TRISO coatings, and fuel compacts) were produced on an engineering scale at BWXT (Lynchburg, Virginia USA) according to the fuel specification (Marshall 2017). This fuel was irradiated in the northeast flux trap (NEFT) at the Advanced Test Reactor (ATR) at Idaho national Laboratory (INL) from February 16, 2018 to July 22, 2020 (Pham et al. 2021). Measurements in the fission product monitoring system (FPMS) indicated unexpected and significant numbers of failures of TRISO particles in Capsule 1 near the end of the sixth irradiation cycle (ATR Cycle 166A). In the fourth cycle (ATR Cycle 164B) and beyond, the sweep gas flow became very low (presumably from degradation of the capsule gas outlet line via an unidentified mechanism), and the program deliberately isolated Capsule 1 from gas flow periodically. In later cycles, attempts to reestablish any kind of flow in Capsule 1 were unsuccessful. With little or no flow through Capsule 1, FPMS measurements and enumerations of failed particles in Capsule 1 were difficult or impossible as was the ability to control the helium/neon gas mixture used for temperature control. Gas flows and fission gas activity in the effluent gas from the other AGR-5/6/7 capsules were also impacted by the Capsule 1 gas flow issues and the large increase in fission gas released from the Capsule 1.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2021

Fast GPU-Based Generation of Large Graph Networks From Degree Distributions

Synthetically generated, large graph networks serve as useful proxies to real-world networks for many graph-based applications. The ability to generate such networks helps overcome several limitations of real-world networks regarding their number, availability, and access. Here, we present the design, implementation, and performance study of a novel network generator that can produce very large graph networks conforming to any desired degree distribution. The generator is designed and implemented for efficient execution on modern graphics processing units (GPUs). Given an array of desired vertex degrees and number of vertices for each desired degree, our algorithm generates the edges of a random graph that satisfies the input degree distribution. Multiple runtime variants are implemented and tested: 1) a uniform static work assignment using a fixed thread launch scheme, 2) a load-balanced static work assignment also with fixed thread launch but with cost-aware task-to-thread mapping, and 3) a dynamic scheme with multiple GPU kernels asynchronously launched from the CPU. The generation is tested on a range of popular networks such as Twitter and Facebook, representing different scales and skews in degree distributions. Results show that, using our algorithm on a single modern GPU (NVIDIA Volta V100), it is possible to generate large-scale graph networks at rates exceeding 50 billion edges per second for a 69 billion-edge network. GPU profiling confirms high utilization and low branching divergence of our implementation from small to large network sizes. For networks with scattered distributions, we provide a coarsening method that further increases the GPU-based generation speed by up to a factor of 4 on tested input networks with over 45 billion edges.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Performance Results on CPU/GPU Exascale Architectures for OMEGA: The Ocean Model for E3SM Global Applications

The US Department of Energy (DOE) conducts climate simulations on some of the world’s largest supercomputers. These exascale machines use heterogeneous architectures with both CPUs and GPUs, and scientific codes must adapt to make full use of this computing power. Los Alamos National Lab is developing Omega: The Ocean Model for E3SM Global Applications, which is specifically designed for modern exascale computers. It uses external libraries that have been optimized for a variety of architectures to run on different supercomputers. Omega is an unstructured-mesh ocean model based on TRiSK numerical methods. It will be the new ocean component of the DOE’s Energy Exascale Earth System Model (E3SM). The algorithms in Omega follow those of the current ocean component, MPAS-Ocean, but it will be written in C++ rather than Fortran to take advantage of the Kokkos performance portability library. Omega spatial operators are written as Kokkos kernels to run efficiently on both CPUs and GPUs. Work on Omega began in 2023 with a new C++ framework for unstructured mesh partitioning, halo exchanges, parallel IO, and Kokkos interfaces. The current version, Omega-0, is being developed to solve the shallow water equations and at present includes all of the tendency terms but not time stepping. Here we share the results of Omega-0 verification and performance testing. Verification includes unit tests implemented with CTest as well as convergence tests in Polaris, an in-house python package with a large suite of test problems. Performance tests compare simulations conducted on CPUs versus GPUs and across different architectures: tests are run on Frontier, which has AMD “Optimized 3rd Gen EPYC” CPUs and AMD MI250X GPUs, as well as Perlmutter, which is composed of AMD EPYC 7763 CPUs and NVIDIA A100 GPUs.

58 GEOSCIENCES↗
DOE OSTI2023

Exascale Multiphysics Nuclear Reactor Simulations for Advanced Designs

ENRICO is a coupled application developed under the U.S. Department of Energy's Exascale Computing Project (ECP) targeting the modeling of advanced nuclear reactors. It couples radiation transport with heat and fluid simulation, including the high-fidelity, highresolution Monte-Carlo code Shift and the Computational fluid dynamics code NekRS. NekRS is a highly-performant open-source code for simulation of incompressible and low-Mach fluid flow, heat transfer, and combustion with a particular focus on turbulent flows in complex domains. It is based on rapidly convergent high-order spectral element discretizations that feature minimal numerical dissipation and dispersion. State-of-the-art multilevel preconditioners, efficient high-order time-splitting methods, and runtime-adaptive communication strategies are built on a fast OCCA-based kernel library, libParanumal, to provide scalability and portability across the spectrum of current and future high-performance computing platforms. On Frontier, Nek5000/RS has recently achieved an unprecedented milestone in breaching over 1 billion spectral elements and 350 billion degrees of freedom. Shift has demonstrated the capability to transport upwards of 1 billion particles per second in full core nuclear reactor simulations featuring complete temperature-dependent, continuous-energy physics on Frontier. Shift achieved a weak-scaling efficiency of 97.8% on 8192 nodes of Frontier and calculated 6 reactions in 214,896 fuel pin regions below 1% statistical error yielding first-of-a-kind resolution for a Monte Carlo transport application.

Hamilton, Steven P.↗
DOE OSTI2024

Machine learning for the redox potential prediction of molecules in organic redox flow battery

Here, organic redox flow batteries (ORFB) are recognized as an innovative technology for the large-scale storage of renewable energy. The redox potential of organic redox-active molecules plays a vital role in their performance. Advanced screening techniques like high-throughput experiment and machine learning (ML) have significantly enhanced organic material performance and transformed the field of ORFB. However, the scarcity of experimental data poses a considerable challenge for ML model development in this domain. In our study, we developed lightweight graph-based Gaussian process regression (GPR) models with GPU-accelerated marginalized graph kernel and hybrid kernel to predict the redox potentials of organic redox-active molecules for ORFBs, specifically focusing on small datasets. To evaluate model accuracy, we created a new experimental database of organic redox-active molecules by the data from hundreds of published papers and assembled previous computational datasets. We also considered some key parameters, such as pH conditions and solvent type, to assess their impact on redox potential prediction. Our GPR model predicted redox potentials with high accuracy across all datasets using minimal training data. The study provides powerful tools for molecule screening and design and delivers valuable guidance on designing training datasets for costly experiments.

25 ENERGY STORAGE↗
DOE OSTI2021

NRC Multiphysics Analysis Capability Deployment FY21: Part 3

This report details the progress and activities of Idaho National Laboratory (INL) on the Nuclear Regulatory Commission (NRC) project “Development and Modeling Support for Advanced Non-Light Water Reactors.” The deliverables completed for this report are: Deliverable 1c: the capability to model gas mixtures was added to Pronghorn. A test problem mimicking the conditions achieved in a depressurized loss of forced cooling (DLOFC) event was solved with both RELAP-5 and Pronghorn. Pronghorn employed a finite vol ume method with the Kurganov-Tadmor discretization. The comparison between the mass fraction spatial profiles computed with RELAP-5 and Pronghorn clearly shows the presence of numerical artifacts (i.e., overly diffusive behavior at low Mach numbers). We confirmed that the problem disappears at higher Mach numbers. We recommend future work on the implementation of a low Mach finite volume formulation to better treat low Mach number problems. Deliverable 2a: we demonstrated two approaches to model the radiation/conduction/natural convection heat transfer across a stagnant gas for the PBMR-400 design using Pronghorn. The first approach is based on the net radiation method, which relies on the computation of view factors with the Multiphysics Object-Oriented Simulation Environment (MOOSE) ray tracing capability. The second method is a traditional thermal resistance approach. The test problems include both 2D and 3D geometries. In all cases, the results show very good agreement during a DLOFC transient. This confirms that the faster thermal resistance method produces solutions that are equivalent to the net radiation method for this geometry. Deliverable 3d: we demonstrated the use of the advection kernel for the delayed neutron precursor equation in Griffin with a 2D MSFR model. The results appear physical but further verification is recommended. We also recommend the addition of conjugate heat transfer to compute the temperatures and model the thermomechanic behavior of the reflectors and other structures. Significant memory and performance issues were encountered in the 3D axisymmetric model. Future work is recommended in this area. Task 8g: this task allows multidimensional MOOSE applications to be coupled to system codes (RELAP-7 and SAM). We implemented a faster multiphysics iteration coupling algorithm, which provides an overall 6× acceleration of the 3D-1D coupling of the core multidi- mensional fluid flow solver and the 1D primary and secondary loop model.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2026

RU-net for automatic characterization of TRISO fuel cross sections

During irradiation, phenomena such as kernel swelling and buffer densification may impact the performance of tristructural isotropic (TRISO) particle fuel. Post-irradiation microscopy is often used to identify these irradiation-induced morphologic changes. However, each fuel compact generally contains thousands of TRISO particles. Manually performing the work to get statistical information on these phenomena is cumbersome and subjective. Here, to reduce the subjectivity inherent in that process and to accelerate data analysis, we used convolutional neural networks (CNNs) to automatically segment cross-sectional images of microscopic TRISO layers. CNNs are a class of machine-learning algorithms specifically designed for processing structured grid data. They have gained popularity in recent years due to their remarkable performance in various computer vision tasks, including image classification, object detection, and image segmentation. In this research, we generated a large irradiated TRISO layer dataset with more than 2,000 microscopic images of cross-sectional TRISO particles and the corresponding annotated images. Based on these annotated images, we used different CNNs to automatically segment different TRISO layers. These CNNs include RU-Net (developed in this study), as well as three existing architectures: U-Net, Residual Network (ResNet), and Attention U-Net. The preliminary results show that the model based on RU-Net performs best in terms of Intersection over Union (IoU). Using CNN models, we can expedite the analysis of TRISO particle cross sections, significantly reducing the manual labor involved and improving the objectivity of the segmentation results.

11 - NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2026

AstraAI v1

AstraAI is an open-source, structure-aware AI coding agent designed for large scientific and DOE-HPC codebases such as AMReX-based applications. Unlike general-purpose coding assistants, AstraAI combines retrieval-augmented generation (RAG) with compiler-level Abstract Syntax Tree (AST) analysis to perform precise, scope-constrained code modifications. It identifies exact function spans, enforces locality of edits, and maintains cross-file invariants, enabling deterministic and build-safe transformations in complex C++/GPU environments. AstraAI is intended for developers working on large, evolving HPC frameworks where correctness, reproducibility, and structural integrity are critical. Typical use cases include modifying physics kernels, updating GPU device lambdas, and performing multi-file refactors without breaking compilation or runtime semantics. Compared to conventional LLM-based coding agents - even those with repository access - AstraAI provides structural guarantees rather than free-form text patches. It minimizes unintended diffs, prevents scope drift, preserves formatting and build stability, and reduces structural hallucinations. By integrating compiler tooling directly into the generation loop, AstraAI transforms AI-assisted coding from probabilistic text editing into deterministic, structure-preserving program transformation suitable for mission-critical scientific software.

Natarajan, Mahesh [Lawrence Berkeley National Labo↗
DOE OSTI2023

Integrate FARM with PID controllers: IES Simulation Ecosystem Control System Development

The FARM (Feasible Actuator Range Modifier) software module is a component of the RAVEN-based FORCE framework for analysis of Integrated Energy Systems (IES). FARM was designed to support the HERON software module in the solution of the optimal dispatch problem for IES units. As the result of HERON-FARM dispatch simulation, the set-point trajectories are optimized to meet constraints on both the production variables (i.e., the variables to be optimized such as the electrical power, the hydrogen production rate, etc.) and the process variables tied to the service life of equipment (e.g., steam flowrate, vessel pressure, turbine firing temperature, etc.) at a coarse time resolution (every 10 or 100 seconds) over long time horizons (several days or weeks). In case the operational constraints need to be met at finer time resolution, the computational burden of HERON-FARM would linearly increase with the sampling rate, and sub-optimal solutions might be obtained. System responses characterized by overshoots and damped oscillations temporarily violating the imposed constraints might occur during abrupt power transients. In this report, a hierarchical control system architecture for the operation of the Dynamic Energy Technology and Integration Laboratory (DETAIL) facility constructed at INL was proposed. First, the preliminary studies on the proposed control strategy for operating the facility and the designed PI controllers were reviewed. In particular, the current approach for generating the set-point trajectories was studied, and its limits were identified. To this aim, the inclusion of a Supervisory Control layer embedding a modified version of the FARM algorithm for preserving the system safe operation over both long and real-time horizons was proposed. In this way, FARM would be applied twice, i.e., the original version (“FARM Validator”) aiding the solution of the power dispatch problem, and the modified version (“FARM Supervisory” coordinating the PI controllers to address the real-time control tasks. Despite the kernel of the two modules is the same algorithm, their roles, tasks, and capabilities are quite different. A detailed description of the role of FARM at addressing low-level control tasks is provided, along with tentative operational procedures for training the models embedded into the algorithm by using the collected experimental data.

42 ENGINEERING↗
DOE OSTI2021

Enabling Scalable and Extensible Memory-mapped Datastores in Userspace

Exascale workloads are expected to incorporate data-intensive processing in close coordination with traditional physics simulations. These emerging scientific, data-analytics and machine learning applications need to access a wide variety of datastores in flat files and structured databases. Programmer productivity is greatly enhanced by mapping datastores into the application process's virtual memory space to provide a unified “in-memory” interface. Currently, memory mapping is provided by system software primarily designed for generality and reliability. However, scalability at high concurrency is a formidable challenge on exascale systems. Also, there is a need for extensibility to support new datastores potentially requiring HPC data transfer services. In this article, we present UMap , a scalable and extensible userspace service for memory-mapping datastores. Furthermore, through decoupled queue management, concurrency aware adaptation, and dynamic load balancing, UMap enables application performance to scale even at high concurrency. We evaluate UMap in data-intensive applications, including sorting, graph traversal, database operations, and metagenomic analytics. Our results show that UMap as a userspace service outperforms an optimized kernel-based service across a wide range of intra-node concurrency by 1.22-1.9 × . We performed two case studies to demonstrate UMap 's extensibility. First, a new datastore residing in remote memory is incorporated into UMap as an application-specific plugin. Second, we present a persistent memory allocator Metall built atop UMap for unified storage/memory.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Nonlinear analog processing with anisotropic nonlinear films

Digital signal processing is the cornerstone of several modern-day technologies, yet in multiple applications it faces critical bottlenecks related to memory and speed constraints. Thanks to recent advances in metasurface design and fabrication, light-based analog computing has emerged as a viable option to partially replace or augment digital approaches. Several light-based analog computing functionalities have been demonstrated using patterned flat optical elements, with great opportunities for integration in compact nanophotonic systems. So far, however, the available operations have been restricted to the linear regime, limiting the impact of this technology to a compactification of Fourier optics systems. In this paper, we introduce nonlinear operations to the field of metasurface-based analog optical processing, demonstrating that nonlinear optical phenomena, combined with nonlocality in flat optics, can be leveraged to synthesize kernels beyond linear Fourier optics, paving the way to a broad range of new opportunities. As a practical demonstration, we report the experimental synthesis of a class of nonlinear operations that can be used to realize broadband, polarization-selective analog-domain edge detection.

analog image processing↗
DOE OSTI2022

Gravothermal evolution of dark matter halos with differential elastic scattering

Here, we study gravothermal evolution of dark matter halos in the presence of differential self-scattering that has strong velocity and angular dependencies. We design controlled N-body simulations to model Rutherford and Møller scatterings in the halo, and follow its evolution in both core-expansion and -collapse phases. The simulations show the commonly-used transfer cross section underestimates the effects of dark matter self-interactions, but the viscosity cross section provides an accurate approximation for modeling angular-dependent dark matter scattering. We investigate thermodynamic properties of the halo, and find that the three moments of the Boltzmann equation under the fluid approximation are satisfied. We further propose a constant effective cross section, which integrates over the halo's characteristic velocity dispersion with weighting kernels motivated by kinetic theory of heat conduction. The effective cross section provides a good approximation to differential self-scattering for most of the halo evolution. It indicates that we can map astrophysical constraints on a constant self-interacting cross section to an SIDM model with velocity- and angular-dependent scatterings.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2021

EGR Dilution and Fuel Property Effects on High-Efficiency Spark-Ignition Flames

Modern spark ignition internal combustion engines rely on fast combustion rates and high dilution to achieve high brake thermal efficiencies. To accomplish this, new engine designs have moved towards increased tumble ratios and stroke-to-bore ratios. Increased tumble ratios correlate positively with increases in turbulent kinetic energy and improved fuel and residual gas mixing, all of which favor faster and more efficient combustion. Longer stroke-to-bore ratios allow higher geometric compression ratios and use of late intake valve closing to control peak compression pressures and temperatures. The addition of dilution to improve efficiency is limited by the resulting increase in combustion instabilities manifested by cycle-to-cycle variability. A number of effects - preferential diffusion, turbulence-combustion interactions, stochastic flow patterns, laminar-turbulent flame kernel transitions, and relative length and velocity scales between flame and turbulence - are believed to be responsible for the increase in cycle-to-cycle variations, where their contributions are likely interlinked. Several studies have shown the influence of stochastic flow characteristics on the nature of combustion instabilities, such as velocity patterns on flame kernel formation and cycle-to-cycle variations in residual gas. However, few have focused on the specific effects of fuel properties. The objective of this work is to contrast the effects of dilution on propane stoichiometric combustion against gasoline. Dilution tolerance experiments were conducted in a purpose-built high stroke-to-bore ratio single cylinder engine with both gasoline and LPG. Three-dimensional full cycle computational fluid dynamics (CFD) simulations employing a level-set combustion approach and Reynolds averaged Navier-Stokes (RANS) turbulence modeling was used to qualitatively assess the changes in length and velocity scales for turbulence and the flame. The experimental results showed that LPG can tolerate higher exhaust gas recirculation (EGR) dilution under a variety of conditions. Analysis of CFD simulations showed that propane flames are likely less sensitive to influences from the flow field due to less thickening of the flame and higher effective flame speeds.

33 ADVANCED PROPULSION SYSTEMS↗
DOE OSTI2026

A Survey on the Expanding Scope and Interdisciplinary Opportunities for Processing-in-Memory Techniques

Processing-in-Memory (PIM) is emerging as a practical path to overcome the limitations of traditional von Neumann architectures. At its core, PIM systems implement computing primitives such as logic operations and multiply-accumulate acceleration through compute-in-memory, near-memory processing, or hybrid designs. The role of memory cells varies widely across technologies, acting as inputs, outputs, or analog accumulators through bit-lines and sense amplifiers. This diversity creates trade-offs in precision, bandwidth, latency, and programmability, making it difficult to build a unified understanding on the progress of the field. In this survey, we organize recent advances of PIM into three areas. First, we discuss the progress on the architectural optimizations of PIM and its integration with both DRAM and emerging non-volatile memories. Second, we examine how PIM is being used to accelerate key computing domains, including generative AI workloads and high-performance kernels, along with new approaches. Third, we highlight the growing adoption of PIM in computational sciences, where it is being applied to solve interdisciplinary problems such as genome analysis, mRNA quantification, mass spectrometry, quantum circuit simulation, wave modeling, and secure computation. Finally, we synthesize the major challenges that continue to slow PIM adoption, including manufacturing constraints, power delivery, thermal reliability, data consistency, runtime and memory-management coordination, and the difficulty of building portable software abstractions without sacrificing commercial viability. This work provides an updated, structured perspective on PIM’s potential across computing and computational sciences and the barriers that must be solved for it to reach its full impact.

Asifuzzaman, Kazi [Oak Ridge National Laboratory (↗
DOE OSTI2024

RAP: Resource-aware Automated GPU Sharing for Multi-GPU Recommendation Model Training and Input Preprocessing

Ensuring high-quality recommendations for newly onboarded users requires the continuous retraining of Deep Learning Recommendation Models (DLRMs) with freshly generated data. To serve the online DLRM retraining, existing solutions use hundreds of CPU computing nodes designated for input preprocessing, causing significant power consumption that surpasses even the power usage of GPU trainers. To this end, we propose RAP, an end-to-end DLRM training framework that supports Resource-aware Automated GPU sharing for DLRM input Preprocessing and Training. The core idea of RAP is to accurately capture the remaining GPU computing resources during DLRM training for input preprocessing, achieving superior training efficiency without requiring additional resources. Specifically, RAP utilizes a co-running cost model to efficiently assess the costs of various input preprocessing operations, and it implements a resource-aware horizontal fusion technique that adaptively merges smaller kernels according to GPU availability, circumventing any interference with DLRM training. In addition, RAP leverages a heuristic searching algorithm that jointly optimizes both the input preprocessing graph mapping and the co-running schedule to maximize the end-to-end DLRM training throughput. The comprehensive evaluation shows that RAP achieves 78.3× speedup on average over CPU-based DLRM input preprocessing frameworks. In addition, the end-to-end training throughput of RAP is only 2.04% lower than the ideal case, which has no input preprocessing overhead.

Wang, Zheng↗
DOE OSTI2022

Microfluidic Uranium Microspheres Production for TRISO and Advanced Fuel Concepts

The purpose of this project was to expand existing internal gelation sol-gel capabilities at PNNL to explore producing uranium dioxide spheres for potential use as fuel kernels in next generation Tri-isotropic (TRISO) particle fuel. This project expands on previous sol-gel efforts at PNNL by (1) increasing the size regime of sphere production from the micro-fluidic range to the milli-fluidic range, and (2) producing uranium spheres. The approach involved first scaling up the channel size of the fluidic system to the millimeter range, with radiation safety considerations in mind; testing and demonstration on non-radioactive surrogate material, cerium oxide; then transitioning to uranium production and finally optimizing system parameters. Commercially available fluidic chips in the desired size range were could not be found, therefore the project designed and fabricated a T-junction with 1mm channels for droplet production. Because the production process is temperature sensitive, prior efforts have involved performing droplet production in a lab freezer. To reduce radiological waste and footprint, two alternative chilling methods were explored using aluminum thermal beads as a chill bath and a custom aluminum block fit to reagent reservoir sizes. Both were successful in the cerium tests, however the aluminum block design outperformed the thermal bead bath and was further adapted for the radiological test and production run in the Radiochemical Processing Laboratory (RPL). Gelation trials were performed to determine an acceptable range of feed solution parameters for the uranium dioxide gels, characterized by R-values, which is determined by the ratio of uranium nitrate to Hexamethylenetetramine (HMTA)/urea in the feed solution. R-values ranging from 1.6-2 were examined in the gelation trials, with only the 1.8 condition being tested in production. The project was successful in demonstrating a proof of concept design for producing uranium dioxide spheres, however further optimization is needed to dial in production parameters and improve sphere quality and homogeneity.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2021

SST-GPU: A Scalable SST GPU Component for Performance Modeling and Profiling

Programmable accelerators have become commonplace in modern computing systems. Advances in programming models and the availability of unprecedented amounts of data have created a space for massively parallel accelerators capable of maintaining context for thousands of concurrent threads resident on-chip. These threads are grouped and interleaved on a cycle-by-cycle basis among several massively parallel computing cores. One path for the design of future supercomputers relies on an ability to model the performance of these massively parallel cores at scale. The SST framework has been proven to scale up to run simulations containing tens of thousands of nodes. A previous report described the initial integration of the open-source, execution-driven GPU simulator, GPGPU-Sim, into the SST framework. This report discusses the results of the integration and how to use the new GPU component in SST. It also provides examples of what it can be used to analyze and a correlation study showing how closely the execution matches that of a Nvidia V100 GPU when running kernels and mini-apps.

97 MATHEMATICS AND COMPUTING↗