Search NASA⌕ Search

SEARCH · Search NASA

Results for “GPU accelerators”

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 667 records · Page 37

SU-F-BRD-02: Application of ARCHERRT-- A GPU-Based Monte Carlo Dose Engine for Radiation Therapy -- to Tomotherapy and Patient-Independent IMRT

Purpose: As a module of ARCHER -- Accelerated Radiation-transport Computations in Heterogeneous EnviRonments, ARCHER{sub RT} is designed for RadioTherapy (RT) dose calculation. This paper describes the application of ARCHERRT on patient-dependent TomoTherapy and patient-independent IMRT. It also conducts a 'fair' comparison of different GPUs and multicore CPU. Methods: The source input used for patient-dependent TomoTherapy is phase space file (PSF) generated from optimized plan. For patient-independent IMRT, the open filed PSF is used for different cases. The intensity modulation is simulated by fluence map. The GEANT4 code is used as benchmark. DVH and gamma index test are employed to evaluate the accuracy of ARCHER{sub RT} code. Some previous studies reported misleading speedups by comparing GPU code with serial CPU code. To perform a fairer comparison, we write multi-thread code with OpenMP to fully exploit computing potential of CPU. The hardware involved in this study are a 6-core Intel E5-2620 CPU and 6 NVIDIA M2090 GPUs, a K20 GPU and a K40 GPU. Results: Dosimetric results from ARCHER{sub RT} and GEANT4 show good agreement. The 2%/2mm gamma test pass rates for different clinical cases are 97.2% to 99.7%. A single M2090 GPU needs 50~79 seconds for the simulation to achieve a statistical error of 1% in the PTV. The K40 card is about 1.7∼1.8 times faster than M2090 card. Using 6 M2090 card, the simulation can be finished in about 10 seconds. For comparison, Intel E5-2620 needs 507∼879 seconds for the same simulation. Conclusion: We successfully applied ARCHER{sub RT} to Tomotherapy and patient-independent IMRT, and conducted a fair comparison between GPU and CPU performance. The ARCHER{sub RT} code is both accurate and efficient and may be used towards clinical applications.

60 APPLIED LIFE SCIENCES↗

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↗

Large-Scale Multi-Dimensional Document Clustering on GPU Clusters

Document clustering plays an important role in data mining systems. Recently, a flocking-based document clustering algorithm has been proposed to solve the problem through simulation resembling the flocking behavior of birds in nature. This method is superior to other clustering algorithms, including k-means, in the sense that the outcome is not sensitive to the initial state. One limitation of this approach is that the algorithmic complexity is inherently quadratic in the number of documents. As a result, execution time becomes a bottleneck with large number of documents. In this paper, we assess the benefits of exploiting the computational power of Beowulf-like clusters equipped with contemporary Graphics Processing Units (GPUs) as a means to significantly reduce the runtime of flocking-based document clustering. Our framework scales up to over one million documents processed simultaneously in a sixteennode GPU cluster. Results are also compared to a four-node cluster with higher-end GPUs. On these clusters, we observe 30X-50X speedups, which demonstrates the potential of GPU clusters to efficiently solve massive data mining problems. Such speedups combined with the scalability potential and accelerator-based parallelization are unique in the domain of document-based data mining, to the best of our knowledge.

99 GENERAL AND MISCELLANEOUS↗

FEL Simulation Using Distributed Computing

While simulation tools are available and have been used regularly for simulating light sources, the increasing availability and lower cost of GPU-based processing opens up new opportunities. This poster highlights a method of how accelerating and parallelizing code processing through the use of COTS software interfaces.

43 PARTICLE ACCELERATORS↗

Celeritas R&D Report: Accelerating Geant4

Celeritas is a new Monte Carlo (MC) detector simulation code designed for computationally intensive applications on high-performance heterogeneous architectures. In the past two years Celeritas has advanced from prototyping a Graphics Processing Unit (GPU)-based single physics model in infinite medium to implementing a full set of electromagnetic (EM) physics processes in complex geometries. The current release of Celeritas, version 0.4, has incorporated full device-based navigation, an event loop in the presence of magnetic fields, and detector hit scoring. New functionality incorporates a scheduler to offload electromagnetic physics to the GPU within a Geant4-driven simulation, enabling straightforward integration of Celeritas into the high energy physics (HEP) experimental frameworks CMSSW and ATLAS FullSimLight. On the Perlmutter supercomputer, Celeritas performs EM physics between 3× and 18× faster using the machine’s Nvidia GPUs compared to using only CPUs, corresponding to an electrical power efficiency up to a factor of 5. When running a multithreaded Geant4 ATLAS test beam application with full hadronic physics, using Celeritas to accelerate the EM physics results in an overall simulation speedup of 1.7–2.2× on GPU and 1.2× on CPU. In a CMS test application using tt¯ events and the prototype Run 4 configuration, compared to Geant4 CPU, Celeritas with a Nvidia A100 improves overall throughput up to a factor of 2.7× but cannot be efficiently shared with more than 8 cores.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Python for Development of OpenMP and CUDA Kernels for Multidimensional Data

Design of data structures for high performance computing (HPC) is one of the principal challenges facing researchers looking to utilize heterogeneous computing machinery. Heterogeneous systems derive cost, power, and speed efficiency by being composed of the appropriate hardware for the task. Yet, each type of processor requires a specific organization of the application state in order to achieve peak performance. Discovering this and refactoring the code can be a challenging and time-consuming task for the researcher, as the data structures and the computational model must be co-designed. We present a methodology that uses Python as the environment for which to explore tradeoffs in both the data structure design as well as the code executing on the computation accelerator. Our method enables multi-dimensional arrays to be used effectively in any target environment. We have chosen to focus on OpenMP and CUDA environments, thus exploring the development of optimized kernels for the two most common classes of computing hardware available today: multi-core CPU and GPU. Python s large palette of file and network access routines, its associative indexing syntax and support for common HPC environments makes it relevant for diverse hardware ranging from laptops through computing clusters to the highest performance supercomputers. Our work enables researchers to accelerate the development of their codes on the computing hardware of their choice.

43 PARTICLE ACCELERATORS↗

Implementing Molecular Dynamics on Hybrid High Performance Computers - Particle-Particle Particle-Mesh

The use of accelerators such as graphics processing units (GPUs) has become popular in scientific computing applications due to their low cost, impressive floating-point capabilities, high memory bandwidth, and low electrical power requirements. Hybrid high-performance computers, machines with nodes containing more than one type of floating-point processor (e.g. CPU and GPU), are now becoming more prevalent due to these advantages. In this paper, we present a continuation of previous work implementing algorithms for using accelerators into the LAMMPS molecular dynamics software for distributed memory parallel hybrid machines. In our previous work, we focused on acceleration for short-range models with an approach intended to harness the processing power of both the accelerator and (multi-core) CPUs. To augment the existing implementations, we present an efficient implementation of long-range electrostatic force calculation for molecular dynamics. Specifically, we present an implementation of the particle-particle particle-mesh method based on the work by Harvey and De Fabritiis. We present benchmark results on the Keeneland InfiniBand GPU cluster. We provide a performance comparison of the same kernels compiled with both CUDA and OpenCL. We discuss limitations to parallel efficiency and future directions for improving performance on hybrid or heterogeneous computers.

97 MATHEMATICS AND COMPUTING↗

Accelerating a 3D Finite-Difference Earthquake Simulation with a C-to-CUDA Translator

GPUs provide impressive computing power, but GPU programming can be challenging. Here, an experience in porting real-world earthquake code to Nvidia GPUs is described. Specifically, an annotation-based programming model, called Mint, and its accompanying source-to-source translator are used to automatically generate CUDA source code and simplify the exploration of performance tradeoffs.

Computer Science↗

Acceleration of tensor-product operations for high-order finite element methods

This article is devoted to graphics processing unit (GPU) kernel optimization and performance analysis of three tensor-product operations arising in finite element methods. We provide a mathematical background to these operations and implementation details. Achieving close to peak performance for these operators requires extensive optimization because of the operators’ properties: low arithmetic intensity, tiered structure, and the need to store intermediate results during the kernel execution. We give a guided overview of optimization strategies and we present a performance model that allows us to compare the efficacy of these optimizations against an empirically calibrated roofline.

Świrydowicz, Kasia↗

JACC.shared: Leveraging HPC Metaprogramming and Performance Portability for Computations That Use Shared Memory GPUs

In this work, we present JACC.shared, a new feature of Julia for ACCelerators (JACC), which is the performanceportable and metaprogramming model of the just-in-time and LLVM-based Julia language. This new feature allows JACC applications to leverage the high-performance computing (HPC) capabilities of high-bandwidth, on-chip GPU memory. Historically, exploiting high-bandwidth, shared-memory GPUs has not been a priority for high-level programming solutions. JACC.shared covers that gap for the first time, thereby providing a highlevel, portable, and easy-to-use solution for programmers to exploit this memory and supporting all current major accelerator architectures. Well-known HPC and AI workloads, such as multi/hyperspectral imaging and AI convolutions, have been used to evaluate JACC.shared on two exascale GPU architectures hosted by some of the most powerful US Department of Energy supercomputers: Perlmutter (NVIDIA A100) and Frontier (AMD MI250X). The performance evaluation reports speedup of up to 3.5× by adding only one line of code to the base codes, thus providing important accelerators in a simple, portable, and transparent way and elevating the programming productivity and performance-portability capabilities for Julia/JACC HPC, AI, and scientific applications.

Valero Lara, Pedro [ORNL] (ORCID:0000000214794310)↗

Bayesian 3D X-ray computed tomography image reconstruction with a scaled Gaussian mixture prior model

In order to improve quality of 3D X-ray tomography reconstruction for Non Destructive Testing (NDT), we investigate in this paper hierarchical Bayesian methods. In NDT, useful prior information on the volume like the limited number of materials or the presence of homogeneous area can be included in the iterative reconstruction algorithms. In hierarchical Bayesian methods, not only the volume is estimated thanks to the prior model of the volume but also the hyper parameters of this prior. This additional complexity in the reconstruction methods when applied to large volumes (from 512{sup 3} to 8192{sup 3} voxels) results in an increasing computational cost. To reduce it, the hierarchical Bayesian methods investigated in this paper lead to an algorithm acceleration by Variational Bayesian Approximation (VBA) [1] and hardware acceleration thanks to projection and back-projection operators paralleled on many core processors like GPU [2]. In this paper, we will consider a Student-t prior on the gradient of the image implemented in a hierarchical way [3, 4, 1]. Operators H (forward or projection) and H{sup t} (adjoint or back-projection) implanted in multi-GPU [2] have been used in this study. Different methods will be evalued on synthetic volume 'Shepp and Logan' in terms of quality and time of reconstruction. We used several simple regularizations of order 1 and order 2. Other prior models also exists [5]. Sometimes for a discrete image, we can do the segmentation and reconstruction at the same time, then the reconstruction can be done with less projections.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

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↗

Implementing Molecular Dynamics on Hybrid High Performance Computers - Three-Body Potentials

The use of coprocessors or accelerators such as graphics processing units (GPUs) has become popular in scientific computing applications due to their low cost, impressive floating-point capabilities, high memory bandwidth, and low electrical power re- quirements. Hybrid high-performance computers, defined as machines with nodes containing more than one type of floating-point processor (e.g. CPU and GPU), are now becoming more prevalent due to these advantages. Although there has been extensive research into methods to efficiently use accelerators to improve the performance of molecular dynamics (MD) employing pairwise potential energy models, little is reported in the literature for models that include many-body effects. 3-body terms are required for many popular potentials such as MEAM, Tersoff, REBO, AIREBO, Stillinger-Weber, Bond-Order Potentials, and others. Because the per-atom simulation times are much higher for models incorporating 3-body terms, there is a clear need for efficient algo- rithms usable on hybrid high performance computers. Here, we report a shared-memory force-decomposition for 3-body potentials that avoids memory conflicts to allow for a deterministic code with substantial performance improvements on hybrid machines. We describe modifications necessary for use in distributed memory MD codes and show results for the simulation of water with Stillinger-Weber on the hybrid Titan supercomputer. We compare performance of the 3-body model to the SPC/E water model when using accelerators. Finally, we demonstrate that our approach can attain a speedup of 5.1 with acceleration on Titan for production simulations to study water droplet freezing on a surface.

Brown, W Michael↗

GraphReduce: Processing Large-Scale Graphs on Accelerator-Based Systems

Recent work on real-world graph analytics has sought to leverage the massive amount of parallelism offered by GPU devices, but challenges remain due to the inherent irregularity of graph algorithms and limitations in GPU-resident memory for storing large graphs. We present GraphReduce, a highly efficient and scalable GPU-based framework that operates on graphs that exceed the device’s internal memory capacity. GraphReduce adopts a combination of edge- and vertex-centric implementations of the Gather-Apply-Scatter programming model and operates on multiple asynchronous GPU streams to fully exploit the high degrees of parallelism in GPUs with efficient graph data movement between the host and device.

Architecture optimization, reuse, performance, ene↗

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↗

Using Hybrid Model OpenSHMEM + CUDA to Implement the SHOC Benchmark Suite

This work describes the process of porting the Scalable HeterOgeneous Computing (SHOC) benchmark suite from the hybrid MPI + CUDA implementation to OpenSHMEM + CUDA. SHOC includes a wide variety of benchmark kernels used to measure accelerator performance in both single node and cluster configurations. The hybrid model implementation attempts to place all major computation on accelerator devices, and uses MPI to synchronize and aggregate results. In some cases, MPI Groups are used to gradually reduce the number of accelerators used for computation as the problem size drops. Porting this behavior to OpenSHMEM required implementing several synchronizing collective operations, and using SHMEM teams to replace MPI Group functionality. Benchmark results on a Cray XK7 system with one GPU per compute node show that SHMEM performance is equal to MPI performance in these hybrid tasks. These results and porting experience show that using OpenSHMEM for accelerator devices benefits from adding functionality for synchronization and teams, and would further benefit from adding support for communication within accelerator kernels.

97 MATHEMATICS AND COMPUTING↗

Portable HCAL reconstruction in the CMS detector using the Alpaka library

CMS has deployed a number of different GPU algorithms at the High-Level Trigger (HLT) in Run 3. As the code base for GPU algorithms continues to grow, the burden for developing and maintaining separate implementations for GPU and CPU becomes increasingly challenging. To mitigate this, CMS has adopted the Alpaka (Abstraction Library for Parallel Kernel Acceleration) library as the performance portability solution to provide a single-code base for parallel execution on both GPUs and CPUs in CMS software (CMSSW). A direct CUDA version of HCAL energy reconstruction, called Minimization At Hcal, Iteratively (MAHI), has been deployed at the HLT in the 2022-2023 data taking period. This contribution will describe how the CUDA version is converted into a portable implementation using the Alpaka library. We will discuss the porting experience from CUDA to Alpaka, the validation process and the performance of the Alpaka version in CPU and GPU.

Kwok, Martin↗

Exploring utilization options of heterogeneous architectures for multi-physics simulations

Top supercomputers in the TOP500 list have transitioned from homogeneous node architectures toward heterogeneous manycore nodes with accelerators and CPUs. These new architectures present significant challenges to developers of large-scale multiphysics applications, especially at DOE laboratories that have invested heavily in scalable MPI codes over decades. Much of these scientific application porting efforts for the new heterogeneous architectures are focused on running the computation on the accelerators, which usually comprise > 90% of the FLOPS of the system, typically using one core per GPU to offload work. Furthermore, we describe an approach to utilizing the remaining FLOPS on a heterogeneous machine by running a portion of the computation on the CPUs cooperatively with the GPU computation. Orthogonally, we evaluate using the additional cores to also offload work to GPUs, effectively oversubscribing the GPUs. We present a proof-of-concept implementation in ARES, a multiphysics ALE-AMR code at LLNL. ARES uses a portability layer, RAJA, which enables us to utilize the same source code for both the CPU and the GPU. We develop an approach to utilize both types of processors cooperatively in a heterogeneous system. Our implementation divides the work between the computing resources via domain decomposition, and utilizes all cores of the CPU and all of the GPUs on the node for computation. Load balancing is necessary to use the heterogeneous resources effectively. We present preliminary results on early delivery pre-Sierra machines at LLNL, showing up to 40% performance benefit of oversubscribing the GPUs (with caveats), and up to an 18% performance benefit of using the CPUs on the heterogeneous nodes for computing in addition to using the GPUs.

97 MATHEMATICS AND COMPUTING↗