Search NASA⌕ Search

SEARCH · Search NASA

Results for “graphic processing units”

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 253 records · Page 14

Automated Hybrid Variance Reduction on Advanced Architectures in the Shift Monte Carlo Code

Monte Carlo transport methods are the most accurate schemes for solving problems with complex energy and spatial features, but they come with a high computational cost. Although hybrid methods have enabled the use of Monte Carlo transport for a large class of problems, they still require significant computing resources. Modern multicore CPUs with large numbers of compute cores and graphical processing units (GPUs) provide opportunities to optimize the memory and run-time costs of hybrid Monte Carlo methods. This paper documents the development and analysis of three Monte Carlo transport algorithms that support hybrid transport using the consistent adjoint-driven importance sampling (CADIS) and forward-weighted CADIS methods in the Shift Monte Carlo code: history-based transport using static and dynamic threading on multicore CPUs and event-based transport enabling weight window tracking on GPUs. The results are shown for two challenging hybrid problems on the Frontier supercomputer at the Oak Ridge Leadership Computing Facility. The results show that all three methods yield good performance and enable solutions of difficult fixed-source transport problems in less than 2 min on 20 nodes of Frontier. Dynamic threading was observed to give up to 20% better scaling behavior than static threading. Moreover, the AMD Instinct 250X GPU was found to give 9 to 11 times greater throughput per graphics compute die than the best CPU performance. In conclusion, additional opportunities for optimization of hybrid transport on GPUs are discussed.

Denovo↗

A Massively Parallel Adaptive Fast-Multipole Method on Heterogeneous Architectures

We present new scalable algorithms and an implementation of the kernel-independent fast multiple method (KIFMM), employing hybrid distributed memory message passing (via MPI) and shared memory/streaming using graphics processing unit (GPU) acceleration to rapidly evaluate two-body non-oscillatory potentials. On traditional CPU-only systems, our implementation scales well up to 30 billion unknowns on 65k cores (AMD/CRAY-based Kraken system at NSF/NICS) on tree data structures with 25 levels between leaves. On GPU-enabled systems, we achieve 30 x speedup for problems of up to 256 million points on 256 GPUs (Lincoln at NSF/NCSA) over a comparable CPU-only implementation. Both of these demonstrations represent the largest and fastest of their kind of which we are aware. We achieve scalability at extreme core counts by extending the initial work of Ying et al. (ACM/IEEE SC 03) with a new approach to scalable MPI-based tree construction and partitioning. For the sub-components of KIFMM, which direct- and approximate-interactions, target evaluation, and source-to-multipole translations, we use CUDA-based GPU-acceleration to achieve excellent performance. To do so requires carefully constructed data structure transformations, which we describe, and whose cost we show is minor. Taken together, these components show promise for ultrascalable FMM in the petascale era and beyond.

99 GENERAL AND MISCELLANEOUS↗

A massively parallel method of characteristic neutral particle transport code for GPUs

Over the past 20 years, parallel computing has enabled computers to grow ever larger and more powerful while scientific applications have advanced in sophistication and resolution. This trend is being challenged, however, as the power consumption for conventional parallel computing architectures has risen to unsustainable levels and memory limitations have come to dominate compute performance. Heterogeneous computing platforms, such as Graphics Processing Units (GPUs), are an increasingly popular paradigm for solving these issues. This paper explores the applicability of GPUs for deterministic neutron transport. A 2D method of characteristics (MOC) code - OpenMOC - has been developed with solvers for both shared memory multi-core platforms as well as GPUs. The multi-threading and memory locality methodologies for the GPU solver are presented. Performance results for the 2D C5G7 benchmark demonstrate 25-35 x speedup for MOC on the GPU. The lessons learned from this case study will provide the basis for further exploration of MOC on GPUs as well as design decisions for hardware vendors exploring technologies for the next generation of machines for scientific computing. (authors)

97 MATHEMATICS AND COMPUTING↗

A Monte Carlo neutron transport code for eigenvalue calculations on a dual-GPU system and CUDA environment

Monte Carlo (MC) method is able to accurately calculate eigenvalues in reactor analysis. Its lengthy computation time can be reduced by general-purpose computing on Graphics Processing Units (GPU), one of the latest parallel computing techniques under development. The method of porting a regular transport code to GPU is usually very straightforward due to the 'embarrassingly parallel' nature of MC code. However, the situation becomes different for eigenvalue calculation in that it will be performed on a generation-by-generation basis and the thread coordination should be explicitly taken care of. This paper presents our effort to develop such a GPU-based MC code in Compute Unified Device Architecture (CUDA) environment. The code is able to perform eigenvalue calculation under simple geometries on a multi-GPU system. The specifics of algorithm design, including thread organization and memory management were described in detail. The original CPU version of the code was tested on an Intel Xeon X5660 2.8 GHz CPU, and the adapted GPU version was tested on NVIDIA Tesla M2090 GPUs. Double-precision floating point format was used throughout the calculation. The result showed that a speedup of 7.0 and 33.3 were obtained for a bare spherical core and a binary slab system respectively. The speedup factor was further increased by a factor of {approx}2 on a dual GPU system. The upper limit of device-level parallelism was analyzed, and a possible method to enhance the thread-level parallelism was proposed. (authors)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Dosimetric evaluation of total marrow irradiation using 2 different planning systems

This study compared 2 different treatment planning systems (TPSs) for quality and efficiency of total marrow irradiation (TMI) plans. The TPSs used in this study were VOxel-Less Optimization (VoLO) (Accuray Inc, Sunnyvale, CA) using helical dose delivery on a Tomotherapy Hi-Art treatment unit and Eclipse (Varian Medical Systems Inc, Palo Alto, CA) using volumetric modulated arc therapy (VMAT) dose delivery on a Varian iX treatment unit. A total dose of 1200 cGy was prescribed to cover 95% of the planning target volume (PTV). The plans were optimized and calculated based on a single CT data and structure set using the Alderson Rando phantom (The Phantom Laboratory, Salem, NY) and physician contoured target and organ at risk (OAR) volumes. The OARs were lungs, heart, liver, kidneys, brain, and small bowel. The plans were evaluated based on plan quality, time to optimize the plan and calculate the dose, and beam on time. The resulting mean and maximum doses to the PTV were 1268 and 1465 cGy for VoLO and 1284 and 1541 cGy for Eclipse, respectively. For 5 of 6 OAR structures the VoLO system achieved lower mean and D10 doses ranging from 22% to 52% and 3% to 44%, respectively. Total computational time including only optimization and dose calculation were 0.9 hours for VoLO and 3.8 hours for Eclipse. These times do not include user-dependent target delineation and field setup. Both planning systems are capable of creating high-quality plans for total marrow irradiation. The VoLO planning system was able to achieve more uniform dose distribution throughout the target volume and steeper dose fall off, resulting in superior OAR sparing. VoLO's graphics processing unit (GPU)–based optimization and dose calculation algorithm also allowed much faster creation of TMI plans.

61 RADIATION PROTECTION AND DOSIMETRY↗

Validation of GPUMCD for low-energy brachytherapy seed dosimetry

Purpose: To validate GPUMCD, a new package for fast Monte Carlo dose calculations based on the GPU (graphics processing unit), as a tool for low-energy single seed brachytherapy dosimetry for specific seed models. As the currently accepted method of dose calculation in low-energy brachytherapy computations relies on severe approximations, a Monte Carlo based approach would result in more accurate dose calculations, taking in to consideration the patient anatomy as well as interseed attenuation. The first step is to evaluate the capability of GPUMCD to reproduce low-energy, single source, brachytherapy calculations which could ultimately result in fast and accurate, Monte Carlo based, brachytherapy dose calculations for routine planning. Methods: A mixed geometry engine was integrated to GPUMCD capable of handling parametric as well as voxelized geometries. In order to evaluate GPUMCD for brachytherapy calculations, several dosimetry parameters were computed and compared to values found in the literature. These parameters, defined by the AAPM Task-Group No. 43, are the radial dose function, the 2D anisotropy function, and the dose rate constant. These three parameters were computed for two different brachytherapy sources: the Amersham OncoSeed 6711 and the Imagyn IsoStar IS-12501. Results: GPUMCD was shown to yield dosimetric parameters similar to those found in the literature. It reproduces radial dose functions to within 1.25% for both sources in the 0.5< r <10 cm range. The 2D anisotropy function was found to be within 3% at r = 5 cm and within 4% at r = 1 cm. The dose rate constants obtained were within the range of other values reported in the literature.Conclusion: GPUMCD was shown to be able to reproduce various TG-43 parameters for two different low-energy brachytherapy sources found in the literature. The next step is to test GPUMCD as a fast clinical Monte Carlo brachytherapy dose calculations with multiple seeds and patient geometry, potentially providing more accurate results than the TG-43 formalism while being much faster than calculations using general purpose Monte Carlo codes.

61 RADIATION PROTECTION AND DOSIMETRY↗

Evaluation of vectorized Monte Carlo algorithms on GPUs for a neutron Eigenvalue problem

Conventional Monte Carlo (MC) methods for radiation transport computations are 'history-based', which means that one particle history at a time is tracked. Simulations based on such methods suffer from thread divergence on the graphics processing unit (GPU), which severely affects the performance of GPUs. To circumvent this limitation, event-based vectorized MC algorithms can be utilized. A versatile software test-bed, called ARCHER - Accelerated Radiation-transport Computations in Heterogeneous Environments - was used for this study. ARCHER facilitates the development and testing of a MC code based on the vectorized MC algorithm implemented on GPUs by using NVIDIA's Compute Unified Device Architecture (CUDA). The ARCHER{sub GPU} code was designed to solve a neutron eigenvalue problem and was tested on a NVIDIA Tesla M2090 Fermi card. We found that although the vectorized MC method significantly reduces the occurrence of divergent branching and enhances the warp execution efficiency, the overall simulation speed is ten times slower than the conventional history-based MC method on GPUs. By analyzing detailed GPU profiling information from ARCHER, we discovered that the main reason was the large amount of global memory transactions, causing severe memory access latency. Several possible solutions to alleviate the memory latency issue are discussed. (authors)

97 MATHEMATICS AND COMPUTING↗

Ultrafast convolution/superposition using tabulated and exponential kernels on GPU

Purpose: Collapsed-cone convolution/superposition (CCCS) dose calculation is the workhorse for IMRT dose calculation. The authors present a novel algorithm for computing CCCS dose on the modern graphic processing unit (GPU). Methods: The GPU algorithm includes a novel TERMA calculation that has no write-conflicts and has linear computation complexity. The CCCS algorithm uses either tabulated or exponential cumulative-cumulative kernels (CCKs) as reported in literature. The authors have demonstrated that the use of exponential kernels can reduce the computation complexity by order of a dimension and achieve excellent accuracy. Special attentions are paid to the unique architecture of GPU, especially the memory accessing pattern, which increases performance by more than tenfold. Results: As a result, the tabulated kernel implementation in GPU is two to three times faster than other GPU implementations reported in literature. The implementation of CCCS showed significant speedup on GPU over single core CPU. On tabulated CCK, speedups as high as 70 are observed; on exponential CCK, speedups as high as 90 are observed. Conclusions: Overall, the GPU algorithm using exponential CCK is 1000-3000 times faster over a highly optimized single-threaded CPU implementation using tabulated CCK, while the dose differences are within 0.5% and 0.5 mm. This ultrafast CCCS algorithm will allow many time-sensitive applications to use accurate dose calculation.

61 RADIATION PROTECTION AND DOSIMETRY↗

Performance study of GPUs in real-time trigger applications for HEP experiments

Graphical Processing Units (GPUs) have evolved into highly parallel, multi-threaded, multicore powerful processors with high memory bandwidth. GPUs are used in a variety of intensive computing applications. The combination of highly parallel architecture and high memory bandwidth makes GPUs a potentially promising technology for effective real-time processing for High Energy Physics (HEP) experiments. However, not much is known of their performance in real-time applications that require low latency, such as the trigger for HEP experiments. We describe an R and D project with the goal to study the performance of GPU technology for possible low latency applications, performing basic operations as well as some more advanced HEP lower-level trigger algorithms (such as fast tracking or jet finding). We present some preliminary results on timing measurements, comparing the performance of a CPU versus a GPU with NVIDIA's CUDA general-purpose parallel computing architecture, carried out at CDF's Level-2 trigger test stand. These studies will provide performance benchmarks for future studies to investigate the potential and limitations of GPUs for real-time applications in HEP experiments.

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↗

A scalable approach to solving dense linear algebra problems on hybrid CPU-GPU systems

Aiming to fully exploit the computing power of all CPUs and all graphics processing units (GPUs) on hybrid CPU-GPU systems to solve dense linear algebra problems, in this paper we design a class of heterogeneous tile algorithms to maximize the degree of parallelism, to minimize the communication volume, and to accommodate the heterogeneity between CPUs and GPUs. The new heterogeneous tile algorithms are executed upon our decentralized dynamic scheduling runtime system, which schedules a task graph dynamically and transfers data between compute nodes automatically. The runtime system uses a new distributed task assignment protocol to solve data dependencies between tasks without any coordination between processing units. By overlapping computation and communication through dynamic scheduling, we are able to attain scalable performance for the double-precision Cholesky factorization and QR factorization. Finally, our approach demonstrates a performance comparable to Intel MKL on shared-memory multicore systems and better performance than both vendor (e.g., Intel MKL) and open source libraries (e.g., StarPU) in the following three environments: heterogeneous clusters with GPUs, conventional clusters without GPUs, and shared-memory systems with multiple GPUs.

97 MATHEMATICS AND COMPUTING↗

Experiences in autotuning matrix multiplication for energy minimization on GPUs

In this study, we report extensive results and analysis of autotuning the computationally intensive graphics processing units kernel for dense matrix–matrix multiplication in double precision. In contrast to traditional autotuning and/or optimization for runtime performance only, we also take the energy efficiency into account. For kernels achieving equal performance, we show significant differences in their energy balance. We also identify the memory throughput as the most influential metric that trades off performance and energy efficiency. Finally, as a result, the performance optimal case ends up not being the most efficient kernel in overall resource use.

97 MATHEMATICS AND COMPUTING↗

Methods for multitasking among real-time embedded compute tasks running on the GPU: Methods for Multitasking Real-time Embedded GPU Computing Tasks

Here, we provide an extensive survey on wide spectrum of scheduling methods for multitasking among graphics processing unit (GPU) computing tasks. We then design several schedulers and explain in detail the selected methods we have developed to implement our scheduling strategies. Next, we compare the performance of schedulers on various workloads running on Fermi and Kepler architectures and arrive at the following major conclusions: (1) Small kernels benefit from running kernels concurrently. (2) The combination of small kernels, high-priority kernels with longer runtimes, and lower-priority kernels with shorter runtimes benefits from a CPU scheduler that dynamically changes kernel order on the Fermi architecture. (3) Because of limitations of existing GPU architectures, currently CPU schedulers outperform their GPU counterparts. We also provide results and observations obtained from implementing and evaluating our schedulers on the NVIDIA Jetson TX1 system-on-chip architecture. We observe that although TX1 has the newer Maxwell architecture, the mechanism used for scheduler timings behaves differently on TX1 compared to Kepler leading to incorrect timings. In this paper, we describe our methods that allow us to report correct timings for CPU schedulers running on TX1. Lastly, we propose new research directions involving the investigation of additional scheduling strategies.

97 MATHEMATICS AND COMPUTING↗

OpenABLext: An automatic code generation framework for agent-based simulations on CPU-GPU-FPGA heterogeneous platforms

The execution of agent-based simulations (ABSs) on hardware accelerator devices such as graphics processing units (GPUs) has been shown to offer great performance potentials. However, in heterogeneous hardware environments, it can become increasingly difficult to find viable partitions of the simulation and provide implementations for different hardware devices. To automate this process, we present OpenABLext, an extension to OpenABL, a model specification language for ABSs. By providing a device-aware OpenCL backend, OpenABLext enables the co-execution of ABS on heterogeneous hardware platforms consisting of central processing units, GPUs, and field programmable gate arrays (FPGAs).We present a novel online dispatching method that efficiently profiles partitions of the simulation during run-time to optimize the hardware assignment while using the profiling results to advance the simulation itself. In addition, OpenABLext features automated conflict resolution based on user-specified rules, supports graph-based simulation spaces, and utilizes an efficient neighbor search algorithm. We show the improved performance of OpenABLext and demonstrate the potential of FPGAs in the context of ABS. We illustrate how co-execution can be used to further lower execution times. OpenABLext can be seen as an enabler to tap the computing power of heterogeneous hardware platforms for ABS.

97 MATHEMATICS AND COMPUTING↗

OpenACC acceleration of an unstructured CFD solver based on a reconstructed discontinuous Galerkin method for compressible flows

Here, an OpenACC directive-based graphics processing unit (GPU) parallel scheme is presented for solving the compressible Navier–Stokes equations on 3D hybrid unstructured grids with a third-order reconstructed discontinuous Galerkin method. The developed scheme requires the minimum code intrusion and algorithm alteration for upgrading a legacy solver with the GPU computing capability at very little extra effort in programming, which leads to a unified and portable code development strategy. A face coloring algorithm is adopted to eliminate the memory contention because of the threading of internal and boundary face integrals. A number of flow problems are presented to verify the implementation of the developed scheme. Timing measurements were obtained by running the resulting GPU code on one Nvidia Tesla K20c GPU card (Nvidia Corporation, Santa Clara, CA, USA) and compared with those obtained by running the equivalent Message Passing Interface (MPI) parallel CPU code on a compute node (consisting of two AMD Opteron 6128 eight-core CPUs (Advanced Micro Devices, Inc., Sunnyvale, CA, USA)). Speedup factors of up to 24× and 1.6× for the GPU code were achieved with respect to one and 16 CPU cores, respectively. The numerical results indicate that this OpenACC-based parallel scheme is an effective and extensible approach to port unstructured high-order CFD solvers to GPU computing.

97 MATHEMATICS AND COMPUTING↗

Vectorization, threading, and cache-blocking considerations for hydrocodes on emerging architectures

This work reports on considerations for improving computational performance in preparation for current and expected changes to computer architecture. The algorithms studied will include increasingly complex prototypes for radiation hydrodynamics codes, such as gradient routines and diffusion matrix assembly (e.g., in [1-6]). The meshes considered for the algorithms are structured or unstructured meshes. The considerations applied for performance improvements are meant to be general in terms of architecture (not specifically graphical processing unit (GPUs) or multi-core machines, for example) and include techniques for vectorization, threading, tiling, and cache blocking. Out of a survey of optimization techniques on applications such as diffusion and hydrodynamics, we make general recommendations with a view toward making these techniques conceptually accessible to the applications code developer. Published 2015. This article is a U.S. Government work and is in the public domain in the USA.

97 MATHEMATICS AND COMPUTING↗

The GPU-enabled divide-expand-consolidate RI-MP2 method (DEC-RI-MP2)

We report porting of the Divide-Expand-Consolidate Resolution of the Identity second-order Møller–Plesset perturbation (DEC-RI-MP2) method to the graphic processing units (GPUs) using OpenACC compiler directives. It is shown that the OpenACC compiler directives implementation efficiently accelerates the rate-determining step of the DEC-RI-MP2 method with minor implementation effort. Moreover, the GPU acceleration results in a better load balance and thus in an overall scaling improvement of the DEC algorithm. The resulting cross-platform hybrid MPI/OpenMP/OpenACC implementation has scalable and portable performance on heterogeneous HPC architectures. The GPU-enabled code was benchmarked using a reduced version of the S12L test set of Stefan Grimme (Grimme, Chem. Eur. J. 2012, 18, 9955) consisting of supramolecular complexes up to 158 atoms and 4292 contracted basis functions (cc-pVTZ). The test set results demonstrate the general applicability of the DEC-RI-MP2 method showing results consistent with the DEC-RI-MP2 introductory paper (Baudin et al., J. Chem. Phys. 2016, 144, 054102) on molecules of complicated electronic structures. © 2016 Wiley Periodicals, Inc.

Chemistry↗

Double-buffered, heterogeneous CPU + GPU integral digestion algorithm for single-excitation calculations involving a large number of excited states

The most widely used quantum-chemical models for excited states are single-excitation theories, a category that includes configuration interaction with single substitutions, time-dependent density functional theory, and also a recently developed ab initio exciton model. When a large number of excited states are desired, these calculations incur a significant bottleneck in the “digestion” step in which two-electron integrals are contracted with density or density-like matrices. We present an implementation that moves this step onto graphical processing units (GPUs), and introduce a double-buffer scheme that minimizes latency by computing integrals on the central processing units (CPUs) concurrently with their digestion on the GPUs. An automatic code generation scheme simplifies the implementation of high-performance GPU kernels. For the exciton model, which requires separate excited-state calculations on each electronically coupled chromophore, the heterogeneous implementation described here results in speedups of 2–6× versus a CPU-only implementation. For traditional time-dependent density functional theory calculations, we obtain speedups of up to 5× when a large number of excited states is computed.

97 MATHEMATICS AND COMPUTING↗