Search NASA⌕ Search

SEARCH · Search NASA

Results for “GPU power”

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 37 records · Page 2
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Advanced Computing is at the Forefront of a New “Moonshot” Revolutionizing the North American Power Grid

In the 50+ years since the first humans landed on the moon, computing has grown at breakneck speed. We are faced with another challenge that is just as daunting, and just as important to overcome-modernizing the North American electric power grid-and high-performance computing (HPC) systems with specialized software will be an important element in rising to this challenge. We describe at a high level how software developed in the ExaSGD project addresses this "moonshot" goal by utilizing exascale computing and a novel high performance solver software stack to support the mission of decarbonizing power grid operations in an environment of uncertain weather and climate. To reach the exascale benchmark the team has made a number of first-of-their-kind innovations, including novel method for stochastic optimization, fine grained parallel methods for modeling power systems, and GPU resident sparse numerical linear solvers.

17 WIND ENERGY↗
DOE OSTI2025

Harnessing distributed GPU computing for generalizable graph convolutional networks in power grid reliability assessments

Although machine learning (ML) has emerged as a powerful tool for rapidly assessing grid contingencies, prior studies have largely considered a static grid topology in their analyses. This limits their application, since they need to be re-trained for every new topology. Here, this paper explores the development of generalizable graph convolutional network (GCN) models by pre-training them across a range of grid topologies and contingency types. We found that a GCN model with auto-regressive moving average (ARMA) layers with a line graph representation of the grid offered the best predictive performance in predicting voltage magnitudes (VM) and voltage angles (VA). We introduced the concept of phantom nodes to consider disparate grid topologies with a varying number of nodes and lines. For pre-training the GCN ARMA model across a variety of topologies, distributed graphics processing unit (GPU) computing afforded us significant training scalability. The predictive performance of this model on grid topologies that were part of the training data is substantially better than the direct current (DC) approximation. Although direct application of the pre-trained model to topologies that are not part of the grid is not particularly satisfactory, fine-tuning with small amounts of data from a specific topology of interest significantly improves predictive performance. In general, this paper highlights the feasibility of training large-scale GNN models to assess the reliability of power grids by considering a wide variety of grid topologies and contingency types. With the advent of foundational models in ML and the exponential increase in GPU computing clusters, generalizable ML models will significantly enhance how utilities manage power systems and make decisions in real-time or near-real-time.

24 - POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2026

Toward an AI-Powered Software Pipeline for Real-Time Tracking and Analysis of Wildfire and Smoke

Real-time tracking of wildfires and smoke is crucial for effective response, minimizing damage, protecting lives, and efficiently managing resources during fire emergencies. We develop a web-based AI-powered pipeline that detects wildfires in aerial video and estimates deployment-relevant behavior metrics, including cumulative burned area, burned-area growth rate, fire spread direction, and smoke dispersion. The system combines a YOLO-based detector with YCbCr-based fire segmentation, HSV-based smoke segmentation, Farneback optical flow, and centroid-based spatiotemporal tracking. Using ground sampling distance (GSD), pixel-level fire masks are converted to physical burned-area measurements by correlating fire pixel counts with camera altitude and tilt angle. We benchmark YOLO variants and non-YOLO baselines (GoogLeNet, CNN, DBN, Autoencoder, U-Net, and AlexNet) on the IEEE FLAME dataset and a newly created aerial frame dataset, Wildfire-DB. Cross-dataset evaluation uses a strict threshold-transfer protocol: decision thresholds are selected on FLAME validation and transferred unchanged to Wildfire-DB to quantify generalization under domain shift. YOLOv6 achieves the strongest cross-dataset frame-level fire detection on Wildfire-DB (ROC-AUC 0.8200, PR-AUC 0.8044, and transferred-threshold F1 0.7596). For tracking-oriented deployment requiring oriented localization, YOLO11-OBB provides the most reliable cross-dataset behavior among OBB-capable models while remaining computationally feasible. To analyze the feasibility of UAV deployment, we further measure inference efficiency using synchronized GPU and CPU power logs on a fixed workload of 1569 frames. YOLO-family models process the video in 5.73–12.47 seconds with net energy of 1247.28–1775.39 J, substantially lower latency and energy than heavier classification and reconstruction baselines. Overall, model optimality depends on operational objectives: YOLOv6 is best for cross-dataset detection robustness, whereas YOL...

Color segmentation↗
DOE OSTI2024

Generic and ML Workloads in an HPC Datacenter: Node Energy, Job Failures, and Node-Job Analysis

HPC datacenters offer a backbone to the modern digital society. Increasingly, they run Machine Learning (ML) jobs next to generic, compute-intensive workloads, supporting science, business, and other decision-making processes. However, understanding how ML jobs impact the operation of HPC datacenters, relative to generic jobs, remains desirable but understudied. In this work, we leverage long-term operational data, collected from a national-scale production HPC datacenter, and statistically compare how ML and generic jobs can impact the performance, failures, resource utilization, and energy consumption of HPC datacenters. Our study provides key insights, e.g., ML-related power usage causes GPU nodes to run into temperature limitations, median/mean runtime and failure rates are higher for ML jobs than for generic jobs, both ML and generic jobs exhibit highly variable arrival processes and resource demands, significant amounts of energy are spent on unsuccessfully terminating jobs, and concurrent jobs tend to terminate in the same state. We open-source our cleaned-up data traces on Zenodo (https://doi. org/10.5281/zenodo.13685426), and provide our analysis toolkit as software hosted on GitHub (https://github.com/atlarge-research/2024-icpads-hpc-workload-characterization). This study offers multiple benefits for data center administrators, who can improve operational efficiency, and for researchers, who can further improve system designs, scheduling techniques, etc.

crossanalysis↗
DOE OSTI2025

NLR HPC Eagle GPU Node Metrics

Ganglia node metrics and iLO (Integrated Lights Out) power data captured from six representative Eagle GPU nodes The Eagle HPC operated at NLR from 2019 through 2024. Eagle was a 2,000-node, 8-petaflop system. This dataset is a representative sample of metrics for 6 of the GPU nodes. Each GPU node contained 2 CPUs and 2 GPUs. Data provided in compressed CSV format. Ganglia and iLO Power Time Series Fields ts: Timestamp dv: Device / Node - Rack and Unit - r103u17 == r(ack)103u(nit)17 mt: Metric (only present for Ganglia) vl: Value - Value in watts for iLO power (instantaneous value at sampling time) or specified Ganglia metric below Ganglia Metrics Metric name -- Metric description -- Unit cpu_aidle -- Percent of time since boot idle CPU -- Percent cpu_idle -- Percent CPU idle -- Percent cpu_nice -- Percent CPU nice -- Percent cpu_speed -- Speed in MHz of CPU -- MHz cpu_user -- Percent CPU user -- Percent cpu_wio -- The percentage of CPU Wait I/O -- Percent gpu0_bar1_memory -- Used GPU bar1 memory -- MB gpu0_decoder_util -- GPU decoder utilization -- Percent gpu0_ecc_db_error -- Total ECC error counts for the GPU -- Number gpu0_encoder_util -- GPU encoder utilization -- Percent gpu0_fan -- Fan speed -- RPM gpu0_fb_memory -- Used GPU framebuffer memory -- MB gpu0_graphics_clock_report -- Current clock speeds for the device -- MHz gpu0_mem_total -- Memory total -- MB gpu0_mem_util -- Memory utilization -- Percent gpu0_power_usage_report -- Power usage report -- Watts gpu0_temp -- GPU 1 temperature -- Celsius gpu1_bar1_memory -- Used GPU bar1 memory -- MB gpu1_decoder_util -- GPU decoder utilization -- Percent gpu1_ecc_db_error -- Total ECC error counts for the GPU -- Number gpu1_encoder_util -- GPU encoder utilization -- Percent gpu1_fan -- Fan speed -- RPM gpu1_fb_memory -- Used GPU framebuffer memory -- MB gpu1_graphics_clock_report -- Current clock speeds for the GPU -- MHz gpu1_mem_total -- Memory total -- MB gpu1_mem_util -- Memory utilization -- MB gpu1_power_usage_report -- Power usage report -- Watts gpu1_temp -- GPU 1 temperature -- Celsius ipmi_cpu1_temp -- CPU 1 temperature -- Celsius ipmi_cpu2_temp -- CPU 2 temperature -- Celsius ipmi_inlet_ambient_temp -- Temperature measured at intake -- Celsius ipmi_vr_p1_temp -- CPU 1 voltage regulator temperature -- Celsius ipmi_vr_p2_temp -- CPU 2 voltage regulator temperature -- Celsius mem_buffers -- Amount of buffered memory -- Bytes mem_cached -- Amount of cached memory -- Bytes mem_free -- Amount of available memory -- Bytes mem_shared -- Amount of shared memory -- Bytes mem_total -- Amount of available memory -- Bytes

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

GPU-resident sparse direct linear solvers for alternating current optimal power flow analysis

Integrating renewable resources within the transmission grid at a wide scale poses significant challenges for economic dispatch as it requires analysis with more optimization parameters, constraints, and sources of uncertainty. This motivates the investigation of more efficient computational methods, especially those for solving the underlying linear systems, which typically take more than half of the overall computation time. In this paper, we present our work on sparse linear solvers that take advantage of hardware accelerators, such as graphical processing units (GPUs), and improve the overall performance when used within economic dispatch computations. We treat the problems as sparse, which allows for faster execution but also makes the implementation of numerical methods more challenging. We present the first GPU-native sparse direct solver that can execute on both AMD and NVIDIA GPUs. We demonstrate significant performance improvements when using high-performance linear solvers within alternating current optimal power flow (ACOPF) analysis. Furthermore, we demonstrate the feasibility of getting significant performance improvements by executing the entire computation on GPU-based hardware. Finally, we identify outstanding research issues and opportunities for even better utilization of heterogeneous systems, including those equipped with GPUs.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2025

FPGA-accelerated SpeckleNN with SNL for real-time X-ray single-particle imaging

We present the implementation of a specialized version of our previously published unified embedding model, SpeckleNN, for real-time speckle pattern classification in X-ray Single-Particle Imaging (SPI), using the SLAC Neural Network Library (SNL) on an FPGA platform. This hardware realization transitions SpeckleNN from a prototypic model into a practical edge solution, optimized for running inference near the detector in high-throughput X-ray free-electron laser (XFEL) facilities, such as those found at the Linac Coherent Light Source (LCLS). To address the resource constraints inherent in FPGAs, we developed a more specialized version of SpeckleNN. The original model, which was designed for broader classification across multiple biological samples, comprised ~5.6 million parameters. The new implementation, while reducing the parameter count to 64.6K (a 98.8% reduction), focuses on maintaining the model's essential functionality for real-time operation, achieving an accuracy of 90%. Furthermore, we compressed the latent space from 128 to 50 dimensions. This implementation was demonstrated on the KCU1500 FPGA board, utilizing 71% of available DSPs, 75% of LUTs, and 48% of FFs, with an average power consumption of 9.4W according to the Vivado post-implementation report. The FPGA performed inference on a single image with a latency of 45.015 microseconds at a 200 MHz clock rate. In comparison, running the same inference on an NVIDIA A100 GPU resulted in an average power consumption of ~73W and an image processing latency of around 400 microseconds. Our FPGA-accelerated version of SpeckleNN demonstrated significant improvements, achieving an 8.9 × speedup and a 7.8 × reduction in power consumption compared to the GPU implementation. Key advancements include model specialization and dynamic weight loading through SNL, which eliminates the need for time-consuming FPGA design re-synthesis, allowing fast and continuous deployment of models (re)trained online. These innovations enable real-time adaptive classification and efficient vetoing of speckle patterns, making SpeckleNN more suited for deployment in XFEL facilities. This implementation has the potential to significantly accelerate SPI experiments and enhance adaptability to evolving experimental conditions.

47 OTHER INSTRUMENTATION↗
DOE OSTI2024

Exploring the Frontiers of Energy Efficiency using Power Management at System Scale

In the face of surging power demands for exascale HPC systems, this work tackles the critical challenge of understanding the impact of software-driven power management techniques like Dynamic Voltage and Frequency Scaling (DVFS) and Power Capping. These techniques have been actively developed over the past few decades. By combining insights from GPU benchmarking to understand application power profiles, we present a telemetry data-driven approach for deriving energy savings projections. This approach has been demonstrably applied to the Frontier supercomputer at scale. Our findings based on three months of telemetry data indicate that, for certain resource-constrained jobs, significant energy savings (up to 8.5%) can be achieved without compromising performance. This translates to a substantial cost reduction, equivalent to 1438 MWh of energy saved. The key contribution of this work lies in the methodology for establishing an upper limit for these best-case scenarios and its successful application. This work enables HPC professionals to optimize the power-performance trade-off within constrained power budgets, not only for the exascale era but also beyond.

Karimi, Ahmad Maroof↗
DOE OSTI2026

Fine-Grained Power and Energy Attribution on AMD GPU/APU-Based Exascale Nodes

Modern exascale GPU- and APU-based systems provide multiple power and energy sensors, but differences in scope, update rate, timing, and filtering complicate the attribution of short-lived accelerator activity. This paper presents a methodology to characterize and correct these effects on Cray EX systems with AMD Instinct MI250X GPUs (Frontier) and MI300A APUs (Portage). Using controlled square-wave workloads, we quantify update intervals, delay, aliasing, and variability across up to 512 GPUs and 480 APUs with on-chip (rocm-smi/amd-smi) and off-chip Cray Power Management sensors. We reconstruct power from cumulative energy counters to achieve faster response times, validate it against on-chip, off-chip, and node-level sensors, and integrate the resulting streams into a Score-P/PAPI-based tool for time-aligned, phase-level attribution. Applied to rocHPL, rocHPL-MxP, and HPG-MxP, the method separates energy savings due to reduced runtime from changes in power. Mixed precision reduces node energy on Frontier by 79% for rocHPL-MxP and 31% for HPG-MxP, with similar trends on Portage. These results provide portable guidance for sensor validation and power-aware optimization on current and future exascale systems.

Mcdaniel, Adam [ORNL] (ORCID:000000016926028X)↗
DOE OSTI2020

CLUE: A Fast Parallel Clustering Algorithm for High Granularity Calorimeters in High-Energy Physics

One of the challenges of high granularity calorimeters, such as that to be built to cover the endcap region in the CMS Phase-2 Upgrade for HL-LHC, is that the large number of channels causes a surge in the computing load when clustering numerous digitized energy deposits (hits) in the reconstruction stage. In this article, we propose a fast and fully parallelizable density-based clustering algorithm, optimized for high-occupancy scenarios, where the number of clusters is much larger than the average number of hits in a cluster. The algorithm uses a grid spatial index for fast querying of neighbors and its timing scales linearly with the number of hits within the range considered. We also show a comparison of the performance on CPU and GPU implementations, demonstrating the power of algorithmic parallelization in the coming era of heterogeneous computing in high-energy physics.

Rovere, Marco↗
DOE OSTI2025

Fine-Grained Application Energy and Power Measurements on the Frontier Exascale System

The increasing complexity and power/energy demands of heterogeneous exascale systems, such as the Frontier supercomputer, present significant challenges for measuring and optimizing power consumption in applications. Current tools either lack the resolution to capture fine-grained power and energy measurements, fail to validate in-band measurements against out-of-band power sensors, or cannot integrate this information with application performance events in a scalable manner. This paper introduces a novel open-source performance toolkit that integrates extended PAPI components with Score-P plugins to enable in-band, fine-grained power and energy measurements, while also supporting validation using power meter measurements for both CPUs and GPUs. One key contribution is the ability to perform millisecond-level power and energy measurements for AMD MI250X GPUs, mapping them to application performance events within a single trace and measurement system that scales. Our toolkit combines coarse-grained measurements from cray_pm counters with high-resolution metrics from rocm_smi and RAPL, converting GPU instantaneous accumulated energy into power to capture both transient and steady-state power behavior, a capability often missed by out-of-band and monitoring tools. By mapping these metrics to specific application regions, developers can identify energy hotspots, address inefficiencies in GPU kernel execution, and validate in-band measurements against external measurements. We demonstrate the effectiveness of this approach through case studies using benchmarks such as GPU rocblas_sgemm, BLIS c_blas_dgemm, and rocHPL, highlighting the variability of the measurements and the impact of transient power spikes on kernel-level efficiency.

Hernandez Mendoza, Oscar [ORNL] (ORCID:00000002538↗
DOE OSTI2020

Pre-exascale accelerated application development: The ORNL Summit experience

High-performance computing (HPC) increasingly relies on heterogeneous architectures to achieve higher performance. In the Oak Ridge Leadership Facility (OLCF), Oak Ridge, TN, USA, this trend continues as its latest supercomputer, Summit, entered production in early 2019. The combination of IBM POWER9 CPU and NVIDIA V100 GPU, along with a fast NVLink2 interconnect and other latest technologies, pushes system performance to a new height and breaks the exascale barrier by certain measures. Due to Summit's powerful GPUs and much higher GPU–CPU ratio, offloading to accelerators becomes a requirement for any application, which intends to effectively use the system. To facilitate navigating a complex landscape of competing heterogeneous architectures, a collection of applications from a wide spectrum of scientific domains is selected for early adoption on Summit. In this article, the experience and lessons learned are summarized, in the hope of providing useful guidance to address new programming challenges, such as scalability, performance portability, and software maintainability, for future application development efforts on heterogeneous HPC systems.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

OpenACC unified programming environment for GPU and FPGA multi-hybrid acceleration

Attached accelerators have been frequently used in recent High Per- formance Computing (HPC) systems because of their high performance/power ratio. In particular, the Graphics Processing Unit (GPU) is the most popu- lar accelerator owing to its high peak FLOPS performance and high memory bandwidth supported by HBM2, etc. However, the performance of GPU depends highly on a large degree of SIMD parallelism and has difficulty sustaining a high performance on programs with frequent branch operations or a partially low degree of parallelism.By contrast, a Field Programmable Gate Array (FPGA) has received attention as a different type of accelerator than GPU as a fully reconfigurable processor fitting the target applications. The high performance of FPGA is mainly provided by a pipelined operation and optimized circuit suitable for any operation even with frequent conditional branches. We have been focusing on the flexibility of FPGA to compensate for the weakness of GPU. We believe that the coupling of GPU with FPGA can result in one of the most powerful accelerating platforms available.However, the program coding of GPU and FPGA coupling can be quite difficult for application users. Traditionally, CUDA by NVIDIA has been the most popular programming language with the largest share of GPUs used in HPC, whereas a hardware description language such as Verilog HDL has been used in FPGA programming. OpenCL coding has recently become available even on high-end FPGAs. Moreover, several recent studies have also enabled the OpenACC coding for use in FPGA. In this study, we provide a unified programming system based on OpenACC for a platform equipped with both GPU and FPGA aiming at the next-generation accelerated supercomputer framework. Our programming environment is called Multi-Hybrid OpenACC Translator (MHOAT), and in this paper, we describe the basic concept and prototype system of MHOAT based on an evaluation on the amount of coding required and the performance of a hybrid multi-device accelerated system.

Tsunashima, Ryuta↗
DOE OSTI2025

Robustness of Deep Learning Classification to Adversarial Input on GPUs: Asynchronous Parallel Accumulation Is a Source of Vulnerability

The ability of machine learning (ML) classification models to resist small, targeted input perturbations—known as adversarial attacks—is a key measure of their safety and reliability. We show that floating-point non associativity (FPNA) coupled with asynchronous parallel programming on GPUs is sufficient to result in misclassification, without any perturbation to the input. Additionally, we show that this misclassification is particularly significant for inputs close to the decision boundary and that standard adversarial robustness results may be overestimated up to 4.6 when not considering machine-level details. We first study a linear classifier, before focusing on standard Graph Neural Network (GNN) architectures and datasets used in robustness assessments. We develop a novel black-box attack using Bayesian optimization to discover external workloads that can change the instruction scheduling which bias the output of reductions on GPUs and reliably lead to misclassification. Motivated by these results, we present a new learnable permutation (LP) gradient-based approach to learning floating-point operation orderings that lead to misclassifications. The LP approach provides a worst-case estimate in a computationally efficient manner, avoiding the need to run identical experiments tens of thousands of times over a potentially large set of possible GPU states or architectures. Finally, using instrumentation-based testing, we investigate parallel reduction ordering across different GPU architectures under external background workloads, when utilizing multi-GPU virtualization, and when applying power capping. Our results demonstrate that parallel reduction ordering varies significantly across architectures under the first two conditions, substantially increasing the search space required to fully test the effects of this parallel scheduler-based vulnerability. These results and the methods developed here can help to include machine-level considerations into adversarial robustness assessments, which can make a difference in safety and mission critical applications.

Shanmugavelu, Sanjif [Maxeler Technologies, a Groq↗
DOE OSTI2024

GPU-enabled extreme-scale turbulence simulations: Fourier pseudo-spectral algorithms at the exascale using OpenMP offloading

Fourier pseudo-spectral methods for nonlinear partial differential equations are of wide interest in many areas of advanced computational science, including direct numerical simulation of three-dimensional (3-D) turbulence governed by the Navier-Stokes equations in fluid dynamics. This paper presents a new capability for simulating turbulence at a new record resolution up to 35 trillion grid points, on the world's first exascale computer, Frontier, comprising AMD MI250x GPUs with HPE's Slingshot interconnect and operated by the US Department of Energy's Oak Ridge Leadership Computing Facility (OLCF). Key programming strategies designed to take maximum advantage of the machine architecture involve performing almost all computations on the GPU which has the same memory capacity as the CPU, performing all-to-all communication among sets of parallel processes directly on the GPU, and targeting GPUs efficiently using OpenMP offloading for intensive number-crunching including 1-D Fast Fourier Transforms (FFT) performed using AMD ROCm library calls. With 99% of computing power on Frontier being on the GPU, leaving the CPU idle leads to a net performance gain via avoiding the overhead of data movement between host and device except when needed for some I/O purposes. Memory footprint including the size of communication buffers for MPI_ALLTOALL is managed carefully to maximize the largest problem size possible for a given node count. Detailed performance data including separate contributions from different categories of operations to the elapsed wall time per step are reported for five grid resolutions, from 2048 3 on a single node to 32768 3 on 4096 or 8192 nodes out of 9408 on the system. Both 1D and 2D domain decompositions which divide a 3D periodic domain into slabs and pencils respectively are implemented. The present code suite (labeled by the acronym GESTS, GPUs for Extreme Scale Turbulence Simulations) achieves a figure of merit (in grid points per second) exceeding goals set in the Center for Accelerated Application Readiness (CAAR) program for Frontier. The performance attained is highly favorable in both weak scaling and strong scaling, with notable departures only for 2048 3 where communication is entirely intra-node, and for 32768 3 , where a challenge due to small message sizes does arise. Communication performance is addressed further using a lightweight test code that performs all-to-all communication in a manner matching the full turbulence simulation code. Performance at large problem sizes is affected by both small message size due to high node counts as well as dragonfly network topology features on the machine, but is consistent with official expectations of sustained performance on Frontier. Overall, although not perfect, the scalability achieved at the extreme problem size of 32768 3 (and up to 8192 nodes — which corresponds to hardware rated at just under 1 exaflop/sec of theoretical peak computational performance) is arguably better than the scalability observed using prior state-of-the-art algorithms on Frontier's predecessor machine (Summit) at OLCF. New science results for the study of intermittency in turbulence enabled by this code and its extensions are to be reported separately in the near future.

3D fast Fourier transform↗
DOE OSTI2020

A massively parallel and scalable multi-CPU material point method

Harnessing the power of modern multi-GPU architectures, we present a massively parallel simulation system based on the Material Point Method (MPM) for simulating physical behaviors of materials undergoing complex topological changes, self-collision, and large deformations. Our system makes three critical contributions. First, we introduce a new particle data structure that promotes coalesced memory access patterns on the GPU and eliminates the need for complex atomic operations on the memory hierarchy when writing particle data to the grid. Second, we propose a kernel fusion approach using a new Grid-to-Particles-to-Grid (G2P2G) scheme, which efficiently reduces GPU kernel launches, improves latency, and significantly reduces the amount of global memory needed to store particle data. Finally, we introduce optimized algorithmic designs that allow for efficient sparse grids in a shared memory context, enabling us to best utilize modern multi-GPU computational platforms for hybrid Lagrangian-Eulerian computational patterns. We demonstrate the effectiveness of our method with extensive benchmarks, evaluations, and dynamic simulations with elastoplasticity, granular media, and fluid dynamics. In comparisons against an open-source and heavily optimized CPU-based MPM codebase [Fang et al. 2019] on an elastic sphere colliding scene with particle counts ranging from 5 to 40 million, our GPU MPM achieves over 100x per-time-step speedup on a workstation with an Intel 8086K CPU and a single Quadro P6000 GPU, exposing exciting possibilities for future MPM simulations in computer graphics and computational science. Moreover, compared to the state-of-the-art GPU MPM method [Hu et al. 2019a], we not only achieve 2x acceleration on a single GPU but our kernel fusion strategy and Array-of-Structs-of-Array (AoSoA) data structure design also generalizes to multi-GPU systems. Our multi-GPU MPM exhibits near-perfect weak and strong scaling with 4 GPUs, enabling performant and large-scale simulations on a 10243 grid with close to 100 million particles with less than 4 minutes per frame on a single 4-GPU workstation and 134 million particles with less than 1 minute per frame on an 8-GPU workstation.

Wang, Xinlei↗