Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “GPU accelerated computing”

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 145 records · Page 8
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Transformational Regional-Scale Earthquake Simulations with the DOE EarthQuake SIMulation Exascale Framework

Earthquakes present worldwide risk to economic and human safety. The 2023 earthquakes in Turkiye provided a reminder of the potential for catastrophic consequences with 50,700 deaths and 15.7 million people affected. The ability to predict ground motions and infrastructure damage for earthquakes continues to be a challenging problem for scientists and engineers. Until now, estimates of ground motions have been performed empirically by looking at sparse data from past earthquakes. This approach can provide statistical information on intensity amplitudes but cannot inform site-specific ground motions essential to developing the most effective resilience. Interest has grown in large-scale computational models to simulate earthquakes at regional scale. The U.S. Department of Energy EarthQuake SIMulation (EQSIM) framework was developed for regional-scale earthquake simulations at unprecedented fidelity, taking advantage of emerging GPU-accelerated systems. This article describes the EQSIM workflow and demonstrates regional-scale simulations with the new computational capability available to scientists in their quest to mitigate future disasters.

58 GEOSCIENCES↗
DOE OSTI2025

ComPort: Rigorous Testing Methods to Safeguard Software Porting (Final Technical Report)

This is a technical report from the lead institution – University of Utah, Kahlert School of Computing – funded under the Department of Energy, Office of Science, Office of Advanced Scientific Computing Research under award number DE-SC0022252. We summarize our work done over the three years of funding received. The relevant papers and software have already been uploaded at the DOE site.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Spy in the GPU-box: Covert and Side Channel Attacks on Multi-GPU System

The deep learning revolution has been enabled in large part by GPUs, and more recently accelerators, which make it possible to carry out computationally demanding training and inference in acceptable times. As the size of machine learning networks and workloads continues to increase, multi-GPU machines have emerged as an important platform offered on High Performance Computing and cloud data centers. Since these machines are shared among multiple users, it becomes increasingly important to protect applications against potential attacks. In this paper, we explore the vulnerability of Nvidia's DGX multi-GPU machines to covert and side channel attacks. These machines consist of a number of discrete GPUs that are interconnected through a combination of custom interconnect (NVLink) and PCIe connections. We reverse engineer the interconnected cache hierarchy and show that it is possible for an attacker on one GPU to cause contention on the L2 cache of another GPU. We use this observation to first develop a covert channel attack across two GPUs, achieving the best bandwidth of around 4 MB/s. We also develop a prime and probe attack on a remote GPU allowing an attacker to recover the cache access pattern of another workload. This access pattern can be used in any number of side channel attacks: we demonstrate a proof of concept attack that fingerprints the application running on the remote GPU, with high accuracy. We also develop a proof of concept attack to extract hyperparameters of a machine learning workload. Our work establishes for the first time the vulnerability of these machines to microarchitectural attacks and can guide future research to improve their security.

Dutta, Sankha↗
DOE OSTI2024

Accelerated Constrained Sparse Tensor Factorization on Massively Parallel Architectures

This study presents the first constrained sparse tensor factorization (cSTF) framework that optimizes and fully offloads computation to massively parallel GPU architectures, and the first performance characterization of cSTF on GPU architectures. In contrast to prior work on tensor factorization, where the matricized tensor times Khatri-Rao product (MTTKRP) is the primary performance bottleneck, our systematic analysis of the cSTF algorithm on GPUs reveals that adding constraints creates an additional bottleneck in the update operation for many real-world sparse tensors. While executing the update operation on the GPU brings significant speedup over its CPU counterpart, it remains a significant bottleneck. To further accelerate the update operation, we propose cuADMM, a new update algorithm that leverages algorithmic and code optimization strategies to minimize both computation and data movement on GPUs. As a result, our framework delivers significantly improved performance compared to prior state-of-the-art. On 10 real-world sparse tensors, our framework achieves geometric mean speedup of 5.1 × (max 41.59 ×) and 7.01 × (max 58.05 ×) on the NIVIDA A100 and H100 GPUs, respectively, over the state-of-the-art SPLATT library running on a 26-core Intel Ice Lake Xeon CPU.

Soh, Yongseok↗
DOE OSTI2025

A Comparison of GPU-Accelerated Multiphase CFD Solvers on the Polaris Supercomputer: Part 1

This report is in support of the Innovative and Novel Computational Impact on Theory and Experiment (INCITE) program sponsored by the U.S. Department of Energy (USDOE). With INCITE-level resources, one project, titled BubblyFlow, was granted computational resources for the 2025 calendar year on the Polaris supercomputer at the Argonne Leadership Computing Facility (ALCF). The project aims to conduct simulations to understand the fundamental characteristics of turbulent bubbly flow phenomena in nature. Staff at the ALCF and Argonne’s Computational Science division, along with collaborators at the City College of New York and University of Illinois at Chicago, helped a summer student to assess the accuracy and performance of two high performance computing (HPC) codes. Both codes, ImExLBM and FluTAS, are fundamentally different in their mathematical and numerical modeling. However, both may be used to solve the same physical problem. The collaboration sought to better understand the differences between both codes in terms of accuracy and efficiency. This would ultimately help the BubblyFlow project better utilize resources and establish a knowledge-base of code capabilities in future simulation campaigns. We compare ImExLBM and FluTAS, two high-performance multiphase computational fluid dynamics (CFD) solvers, in terms of physical fidelity, time-to-solution, and parallel efficiency. We validate ImExLBM (Implicit-Explicit Lattice Boltzmann Method) against a canonical benchmark and assess it’s performance relative to FluTAS (Fluid Transport Accelerated Solver), a well-established open-source CFD code.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Parallel quantum computing simulations via quantum accelerator platform virtualization

Quantum circuit execution is a central task in quantum computation. Due to inherent quantum-mechanical constraints, quantum computing workflows often involve a considerable number of independent measurements over a large set of slightly different quantum circuits. Here we discuss a simple model for parallelizing such quantum circuit executions that is based on introducing a large array of virtual quantum processing units (mapped to HPC nodes in our case) as a parallel quantum computing platform. Implemented within the XACC framework, the model can readily take advantage of its backend-agnostic features, enabling parallel quantum computing/simulation over any target backend supported by XACC. We illustrate the performance of this approach by demonstrating strong scaling in two pertinent domain science problems, namely in computing the gradients for the multi-contracted variational quantum eigensolver and in data-driven quantum circuit learning, where we vary the number of qubits and the number of circuit layers. Here, the latter simulation leverages the cuQuantum library to run efficiently on GPU-accelerated HPC platforms.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Performance Analysis of PIConGPU: Particle-in-Cell on GPUs using NVIDIA’s NSight Systems and NSight Compute

PIConGPU, Particle In Cell on GPUs, is an open source simulations framework for plasma and laser-plasma physics used to develop advanced particle accelerators for radiation therapy of cancer, high energy physics and photon science. While PIConGPU has been optimized for at least 5 years to run well on NVIDIA GPU-based clusters, there has been limited exploration by the development team of potential scalability bottlenecks using recently updated and new tools including NVIDIA’s NVProf tool and the brand-new NVIDIA NSight Suite (Systems and Compute) tools. PIConGPU is a highly optimized application that runs production jobs at scale on a system Oak Ridge Leadership Facility’s (OLCF) Summit supercomputer (using the full machine at 4600 nodes; at 98% of GPU utilization on all ~28000 NVIDIA Volta GPUs). PIConGPU has been selected as one of the the eight applications for OLCF’s coveted Center for Accelerated Application Readiness (CAAR) program aimed at the facility’s Frontier supercomputer (OLCF’s first exascale system to launch in 2021), to partner with our vendors (primary vendors: AMD and Cray/HPE) ensuring that Frontier will be able to perform large-scale science when it opens to users in 2022. To this effect, performance engineers on the PIConGPU team wanted to dive deep into the application to understand at the finest granularity, which portions of the code could be further optimized to exploit the hardware on Summit at it’s maximum potential and also to elucidate which key kernels should be tracked and optimized for the CAAR effort to port this code to Frontier. Any bottlenecks that are observed via performance profiling on Summit are likely to also impact scalability on the Frontier-dev system and the Frontier Early Access (EA) system. Additionally, the engineers wanted to take a closer look at the newest NVIDIA profiling tools which allows us to identify the most useful features on these tools and will provide an opportunity to compare it to new AMD and Cray’s performance analysis tool releases and provide feedback to our vendor partners on what features are most important and mission critical for CAAR efforts. The primary goal of this report is to focus on the evaluation of PIConGPU’s most time-intensive kernels using NVProf and NSight Suite. Three kernels, Current Deposition (also known as Compute Current), Particle Push (Move and Mark), and Shift Particles are known to be some of the most time-consuming kernels in PIConGPU. The Current xi Deposition kernel and Particle Push kernel both set up the particle attributes for running any physics simulation with PIConGPU, so it is crucial to improve the performance of these two kernels. In this report, we measure single GPU metrics for the three kernels, offer high level takeaways from the conducted analysis, and compare the profiling data from NSight Compute to that of NVProf. This analysis was performed using a grid size of 240 x 272 x 224, and 10 time steps with the Mid-November Figure of Merit (FOM) run setup. The Traveling Wave Electron Acceleration (TWEAC) science case used in this run is a representative science case for PIConGPU. This execution can also be used for baseline analysis on AMD MI50/ MI60 systems. As of the time of writing, the PIConGPU application has limited use for features of NSight Systems, so this report will mainly focus on insights garnered from NSight Compute. For this analysis, we run the “full” metric set available in NSight Compute version 2020.1.2 and use NSight Systems version 2020.3.1 to generate the application timeline.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2022

Developments in Performance and Portability for MadGraph5_aMC@NLO

Event generators simulate particle interactions using Monte Carlo techniques, providing the primary connection between experiment and theory in experimental high energy physics. These software packages, which are the first step in the simulation worflow of collider experiments, represent approximately 5 to 20% of the annual WLCG usage for the ATLAS and CMS experiments. With computing architectures becoming more heterogeneous, it is important to ensure that these key software frameworks can be run on future systems, large and small. In this contribution, recent progress on porting and speeding up the Madgraph5_aMC@NLO event generator on hybrid architectures, i.e. CPU with GPU accelerators, is discussed. The main focus of this work has been in the calculation of scattering amplitudes and "matrix elements", which is the computational bottleneck of an event generation application. For physics processes limited to QCD leading order, the code generation toolkit has been expanded to produce matrix element calculations using C++ vector instructions on CPUs and using CUDA for NVidia GPUs, as well as using Alpaka, Kokkos and SYCL for multiple CPU and GPU architectures. Performance is reported in terms of matrix element calculations per time on NVidia, Intel, and AMD devices. The status and outlook for the integration of this work into a production release usable by the LHC experiments, with the same functionalities and very similar user interfaces as the current Fortran version, is also described.

Valassi, Andrea↗
DOE OSTI2023

A survey of techniques for optimizing transformer inference

Recent years have seen a phenomenal rise in the performance and applications of transformer neural networks. The family of transformer networks, including Bidirectional Encoder Representations from Transformer (BERT), Generative Pretrained Transformer (GPT) and Vision Transformer (ViT), have shown their effectiveness across Natural Language Processing (NLP) and Computer Vision (CV) domains. Transformer-based networks such as ChatGPT have impacted the lives of common men. However, the quest for high predictive performance has led to an exponential increase in transformers' memory and compute footprint. Researchers have proposed techniques to optimize transformer inference at all levels of abstraction. Further, this paper presents a comprehensive survey of techniques for optimizing the inference phase of transformer networks. We survey techniques such as knowledge distillation, pruning, quantization, neural architecture search and lightweight network design at the algorithmic level. We further review hardware-level optimization techniques and the design of novel hardware accelerators for transformers. We summarize the quantitative results on the number of parameters/FLOPs and the accuracy of several models/techniques to showcase the tradeoff exercised by them. We also outline future directions in this rapidly evolving field of research. We believe that this survey will educate both novice and seasoned researchers and also spark a plethora of research efforts in this field.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

High-order matrix-free incompressible flow solvers with GPU acceleration and low-order refined preconditioners

In this work, we present a matrix-free flow solver for high-order finite element discretizations of the incompressible Navier-Stokes and Stokes equations with GPU acceleration. For high polynomial degrees, assembling the matrix for the linear systems resulting from the finite element discretization can be prohibitively expensive, both in terms of computational complexity and memory. For this reason, it is necessary to develop matrix-free operators and preconditioners, which can be used to efficiently solve these linear systems without access to the matrix entries themselves. The matrix-free operator evaluations utilize GPU-accelerated sum-factorization techniques to minimize memory movement and maximize throughput. The preconditioners developed in this work are based on a low-order refined methodology with parallel subspace corrections, as described for diffusion problems in [1]. The saddle-point Stokes system is solved using block-preconditioning techniques, which are robust in mesh size, polynomial degree, time step, and viscosity. For the incompressible Navier-Stokes equations, we make use of projection (fractional step) methods, which require Helmholtz and Poisson solves at each time step. The performance of our flow solvers is assessed on several benchmark problems in two and three spatial dimensions.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Ipopt Interface to Re::Solve Linear Solver

The software provides Ipopt optimization package an interface to the Re::Solve linear solver library. Re::Solve features GPU-resident direct and iterative linear solvers that could be used to accelerate optimization computations.

Alam, Maksudul [Oak Ridge National Laboratory (ORN↗
DOE OSTI2025

GPU Acceleration in SRW: Design and Considerations

Synchrotron Radiation Workshop (SRW) is a powerful tool for simulation synchrotron radiation emission and propagation through beamline elements, enabling advanced beamline design and experimental optimization. Recently, GPU acceleration has been developed for SRW to support highly detailed end-to-end simulations of experiments at synchrotron light sources. This work documents the design and implementation of this GPU acceleration support, addressing the complexities of adapting CPU-based components to heterogeneous computing architectures.

43 PARTICLE ACCELERATORS↗
DOE OSTI2024

Assessing Accelerator Library Integration in MOOSE

MOOSE is a general purpose open source multiphysics framework supporting native finite element and finite volume discretizations as well and wrapping other libraries providing arbitrary computational capabilities. Due to its generality, it has experienced significant success. However, with recent changes in the landscape of computer architectures, most notably the growth of GPU computing, MOOSE must assess new technologies or else risk alienating customers interested in the benefits these technologies can offer. In that vein we have assessed multiple accelerator libraries developed through the ECP project, including Kokkos, libCEED, and MFEM, and present our evaluation of these libraries as candidates for incorporation into the MOOSE framework.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

TorchBraid: High-Performance Layer-Parallel Training of Deep Neural Networks with MPI and GPU Acceleration

TorchBraid is a high-performance implementation of layer-parallel training for deep neural networks (DNNs) supporting MPI-based parallelism and GPU acceleration. Layer-parallel training has been developed to overcome the serialization inherent in forward and backward propagation of DNNs that limits utilization of computational resources in the strong scaling limit. To achieve this, TorchBraid integrates the PyTorch neural network framework with the state-of-the-art XBraid time-parallel library. Furthermore, this article presents the use and performance of TorchBraid, in addition to solutions for overcoming the algorithmic challenges inherent in combining automatic differentiation with layer-parallel. Results are presented with and without GPU acceleration for the Tiny ImageNet and MNIST image classification data sets, as well as recurrent neural networks. Overall, TorchBraid enables fast training of DNNs, both in a strong and weak scaling context. In addition to the TorchBraid software, several new advances in applying layer-parallel algorithms are detailed. Integration of layer-parallel with data-parallel algorithms is presented for the first time, showing the computational advantages of the combination. Standard deep learning techniques, like batch-normalization, are developed for layer-parallel training. Finally, a new approach combining layer-parallel with spatial coarsening in order to accelerate training for 3D image classification shows roughly a 10× speedup over serial execution.

Layer-parallel↗
DOE OSTI2022

Scalable Gaussian Processes, GPyTorch Application Benchmarking, and Targeted Adaptive Design (TAD) on ThetaGPU

We aim at showcasing the scalability of Gaussian Process (GP). The naive GP implementation scales cubically with data size, which can be prohibitive, so GP has not heretofore been considered suitable for very large-scale problem settings. We take advantage of GPyTorch, a library for scalable GPs built on top of PyTorch that incorporates GPU acceleration. With GPyTorch, one can achieve nearly linear scaling with structured kernel interpolation (SKI) and constant-time predictive covariances computation with LanczOs Variance Estimates (LOVE) while preserving accuracy. We also take advantage of the computational power of ThetaGPU, a supercomputer of Argonne Leadership Computing Facility (ALCF). In addition, we implement a scalable, GPU-ready version of Targeted Adaptive Design (TAD), a GP-based data-driven algorithm that efficiently searches the control space of an advanced manufacturing experiment for settings capable of producing a required design within a specified tolerance, despite the poorly known mapping from control settings to design. We finally show our benchmarking for GPyTorch and TAD performance on CPU vs. ThetaGPU and discuss the results and implications.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

DUNE Software and High Performance Computing

DUNE, like other HEP experiments, faces a challenge related to matching execution patterns of our production simulation and data processing software to the limitations imposed by modern high-performance computing facilities. In order to efficiently exploit these new architectures, particularly those with high CPU core counts and GPU accelerators, our existing software execution models require adaptation. In addition, the large size of individual units of raw data from the far detector modules pose an additional challenge somewhat unique to DUNE. Here we describe some of these problems and how we begin to solve them today with existing software frameworks and toolkits. We also describe ways we may leverage these existing software architectures to attack remaining problems going forward. This whitepaper is a contribution to the Computational Frontier of Snowmass21.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2023

Scaled ILU Smoothers for Navier-Stokes Pressure Projection

Incomplete LU (ILU) smoothers are effective in the algebraic multigrid (AMG) V-cycle for reducing high-frequency components of the error. However, the requisite direct triangular solves are comparatively slow on GPUs. Previous work has demonstrated the advantages of Jacobi iteration as an alternative to direct solution of these systems. Depending on the threshold and fill-level parameters chosen, the factors can be highly nonnormal and Jacobi is unlikely to converge in a low number of iterations. We demonstrate that row scaling can reduce the departure from normality, allowing us to replace the inherently sequential solve with a rapidly converging Richardson iteration. There are several advantages beyond the lower compute time. Scaling is performed locally for a diagonal block of the global matrix because it is applied directly to the factor. Further, an ILUT Schur complement smoother maintains a constant GMRES iteration count as the number of MPI ranks increases, and thus parallel strong-scaling is improved. Our algorithms have been incorporated into hypre, and we demonstrate improved time to solution for linear systems arising in the Nalu-Wind and PeleLM pressure solvers. For large problem sizes, GMRES+AMG executes at least five times faster when using iterative triangular solves compared with direct solves on massively parallel GPUs.

algebraic multigrid↗
DOE OSTI2023

DGaaS: GPU as a Service on Distributed Computing System

In the rapidly evolving landscape of scientific computing, Graphics Processing Units (GPUs) have become indispensable for their unparalleled ability to handle parallel tasks in complex calculations, simulations, and data analysis. Their utility is further magnified in machine learning and AI applications, where they significantly accelerate model training and predictive analytics. Within this context, the Triton Inference Server emerges as a pivotal open-source tool, specializing in AI inferencing and optimizing GPU utilization across various platforms and frameworks. This paper presents an in-depth study on distributed High Throughput Computing (HTC), specifically focusing on the HTCondor framework and its resource provisioning tools, GlideinWMS and HEPCloud. These systems enable large-scale scientific experiments like CMS and DUNE to efficiently access and utilize vast computational resources. The paper explores the core architectural components of GlideinWMS, including jobs, user pools, and worker nodes, and discusses their integration with GPUs and the Triton server. The primary aim of this research is to develop a solution that optimizes GPU utilization by leveraging Glideins and containers. This approach allows computational jobs, particularly those involving AI models, to use GPUs only when essential, thereby facilitating efficient sharing of limited GPU resources. To validate this architecture, the study conducted three key tests involving custom scripts, container-based servers, and Triton server deployments. However, the study faces challenges, notably in locating the Triton server and ensuring secure remote access. To address these issues, future work will focus on developing a proxy mechanism and enhancing security protocols. In conclusion, this study offers a comprehensive roadmap for effective and efficient GPU utilization in distributed High Throughput Computing. It aims to contribute significantly to the scientific community by solving pressing problems and implementing robust solutions in collaboration with the GlideinWMS and HEPCloud teams. The research sets the stage for a more efficient, scalable, and cost-effective paradigm in scientific computing.

97 MATHEMATICS AND COMPUTING↗