Search NASA⌕ Search

SEARCH · Search NASA

Results for “Speedup”

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 109 records · Page 6

High-throughput virtual laboratory for drug discovery using massive datasets

Time-to-solution for structure-based screening of massive chemical databases for COVID-19 drug discovery has been decreased by an order of magnitude, and a virtual laboratory has been deployed at scale on up to 27,612 GPUs on the Summit supercomputer, allowing an average molecular docking of 19,028 compounds per second. Over one billion compounds were docked to two SARS-CoV-2 protein structures with full optimization of ligand position and 20 poses per docking, each in under 24 hours. GPU acceleration and high-throughput optimizations of the docking program produced 350× mean speedup over the CPU version (50× speedup per node). GPU acceleration of both feature calculation for machine-learning based scoring and distributed database queries reduced processing of the 2.4 TB output by orders of magnitude. The resulting 50× speedup for the full pipeline reduces an initial 43 day runtime to 21 hours per protein for providing high-scoring compounds to experimental collaborators for validation assays.

97 MATHEMATICS AND COMPUTING↗

ConnectIt: a framework for static and incremental parallel graph connectivity algorithms

Connected components is a fundamental kernel in graph applications. The fastest existing multicore algorithms for solving graph connectivity are based on some form of edge sampling and/or linking and compressing trees. However, many combinations of these design choices have been left unexplored. In this paper, we design the ConnectIt framework, which provides different sampling strategies as well as various tree linking and compression schemes. ConnectIt enables us to obtain several hundred new variants of connectivity algorithms, most of which extend to computing spanning forest. In addition to static graphs, we also extend ConnectIt to support mixes of insertions and connectivity queries in the concurrent setting. We present an experimental evaluation of ConnectIt on a 72-core machine, which we believe is the most comprehensive evaluation of parallel connectivity algorithms to date. Compared to a collection of state-of-the-art static multicore algorithms, we obtain an average speedup of 12.4x (2.36x average speedup over the fastest existing implementation for each graph). Using ConnectIt, we are able to compute connectivity on the largest publicly-available graph (with over 3.5 billion vertices and 128 billion edges) in under 10 seconds using a 72-core machine, providing a 3.1x speedup over the fastest existing connectivity result for this graph, in any computational setting. For our incremental algorithms, we show that our algorithms can ingest graph updates at up to several billion edges per second. To guide the user in selecting the best variants in ConnectIt for different situations, we provide a detailed analysis of the different strategies. Finally, we show how the techniques in ConnectIt can be used to speed up two important graph applications: approximate minimum spanning forest and SCAN clustering.

Computer Science↗

Optimizing the performance and structure of the D0 Collie confidence limit evaluator

D0 Collie is a program used to perform limit calculations based on ensembles of pseudo-experiments ('PEs'). Since the application of this program to the crucial Higgs mass limit is quite CPU intensive, it has been deemed important to carefully review this program, with an eye toward identifying and implementing potential performance improvements. At the same time, we identify any coding errors or opportunities for potential structural (or algorithm) improvement discovered in the course of gaining sufficient understanding of the workings of Collie to sensibly explore for optimizations. Based on a careful analysis of the program, a series of code changes with potential for improving performance has been identified. The implementation and evaluation of the most important parts of this series has been done, with gratifying speedup results. The bottom line: We have identified and implemented changes leading to a factor of 2.19 speedup in the example program provided, and expected to translate to a factor of roughly 4 speedup in typical realistic usage.

99 GENERAL AND MISCELLANEOUS↗

Modeling laser wakefield accelerators in a Lorentz boosted frame

Modeling of laser-plasma wakefield accelerators in an optimal frame of reference [1] is shown to produce orders of magnitude speed-up of calculations from first principles. Obtaining these speedups requires mitigation of a high frequency instability that otherwise limits effectiveness in addition to solutions for handling data input and output in a relativistically boosted frame of reference. The observed high-frequency instability is mitigated using methods including an electromagnetic solver with tunable coefficients, its extension to accomodate Perfectly Matched Layers and Friedman's damping algorithms, as well as an efficient large bandwidth digital filter. It is shown that choosing theframe of the wake as the frame of reference allows for higher levels of filtering and damping than is possible in other frames for the same accuracy. Detailed testing also revealed serendipitously the existence of a singular time step at which the instability level is minimized, independently of numerical dispersion, thus indicating that the observed instability may not be due primarily to Numerical Cerenkov as has been conjectured. The techniques developed for Cerenkov mitigation prove nonetheless to be very efficient at controlling the instability. Using these techniques, agreement at the percentage level is demonstrated between simulations using different frames of reference, with speedups reaching two orders of magnitude for a 0.1 GeV class stages. The method then allows direct and efficient full-scale modeling of deeply depleted laser-plasma stages of 10 GeV-1 TeV for the first time, verifying the scaling of plasma accelerators to very high energies. Over 4, 5 and 6 orders of magnitude speedup is achieved for the modeling of 10 GeV, 100 GeV and 1 TeV class stages, respectively.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Optimization of Asynchronous Communication Operations through Eager Notifications

UPC++ is a C++ library implementing the Asynchronous Partitioned Global Address Space (APGAS) model. We propose an enhancement to the completion mechanisms of UPC++ used to synchronize communication operations that is designed to reduce overhead for on-node operations. Our enhancement permits eager delivery of completion notification in cases where the data transfer semantics of an operation happen to complete synchronously, for example due to the use of shared-memory bypass. This semantic relaxation allows removing significant overhead from the critical path of the implementation in such cases. We evaluate our results on three different representative systems using a combination of microbenchmarks and five variations of the the HPCChallenge RandomAccess benchmark implemented in UPC++ and run on a single node to accentuate the impact of locality. We find that in RMA versions of the benchmark written in a straightforward manner (without manually optimizing for locality), the new eager notification mode can provide up to a 25% speedup when synchronizing with promises and up to a 13.5x speedup when synchronizing with conjoined futures. We also evaluate our results using a graph matching application written with UPC++ RMA communication, where we measure overall speedups of as much as 11% in single-node runs of the unmodified application code, due to our transparent enhancements.

Kamil, Amir↗

Thermal Radiation Transport with Tensor Trains

We present a novel tensor network algorithm to solve the time-dependent, gray thermal radiation transport equation. The method invokes a tensor train (TT) decomposition for the specific intensity. The efficiency of this approach is dictated by the rank of the decomposition. When the solution is “low rank,” the memory footprint of the specific intensity solution vector may be significantly compressed. The algorithm, following a step-then-truncate approach of a traditional discrete ordinates method, operates directly on the compressed state vector, thereby enabling large speedups for low-rank solutions. To achieve these speedups, we rely on a recently developed rounding approach based on the Gram-SVD. We detail how familiar S N algorithms for (gray) thermal transport can be mapped to this TT framework and present several numerical examples testing both the optically thick and thin regimes. The TT framework finds low-rank structure and supplies up to ≃60× speedups and ≃1000× compressions for problems demanding large angle counts, thereby enabling previously intractable SN calculations and supplying a promising avenue to mitigate ray effects.

79 ASTRONOMY AND ASTROPHYSICS↗

Porting the WAVEWATCH III (v6.07) wave action source terms to GPU

Abstract. Surface gravity waves play a critical role in several processes, including mixing, coastal inundation, and surface fluxes. Despite the growing literature on the importance of ocean surface waves, wind–wave processes have traditionally been excluded from Earth system models (ESMs) due to the high computational costs of running spectral wave models. The development of the Next Generation Ocean Model for the DOE’s (Department of Energy) E3SM (Energy Exascale Earth System Model) Project partly focuses on the inclusion of a wave model, WAVEWATCH III (WW3), into E3SM. WW3, which was originally developed for operational wave forecasting, needs to be computationally less expensive before it can be integrated into ESMs. To accomplish this, we take advantage of heterogeneous architectures at DOE leadership computing facilities and the increasing computing power of general-purpose graphics processing units (GPUs). This paper identifies the wave action source terms, W3SRCEMD, as the most computationally intensive module in WW3 and then accelerates them via GPU. Our experiments on two computing platforms, Kodiak (P100 GPU and Intel(R) Xeon(R) central processing unit, CPU, E5-2695 v4) and Summit (V100 GPU and IBM POWER9 CPU) show respective average speedups of 2× and 4× when mapping one Message Passing Interface (MPI) per GPU. An average speedup of 1.4× was achieved using all 42 CPU cores and 6 GPUs on a Summit node (with 7 MPI ranks per GPU). However, the GPU speedup over the 42 CPU cores remains relatively unchanged (∼ 1.3×) even when using 4 MPI ranks per GPU (24 ranks in total) and 3 MPI ranks per GPU (18 ranks in total). This corresponds to a 35 %–40 % decrease in both simulation time and usage of resources. Due to too many local scalars and arrays in the W3SRCEMD subroutine and the huge WW3 memory requirement, GPU performance is currently limited by the data transfer bandwidth between the CPU and the GPU. Ideally, OpenACC routine directives could be used to further improve performance. However, W3SRCEMD would require significant code refactoring to make this possible. We also discuss how the trade-off between the occupancy, register, and latency affects the GPU performance of WW3.

58 GEOSCIENCES↗

Computational efficiency of parallel combinatorial OR-tree searches

The performance of parallel combinatorial OR-tree searches is analytically evaluated. This performance depends on the complexity of the problem to be solved, the error allowance function, the dominance relation, and the search strategies. The exact performance may be difficult to predict due to the nondeterminism and anomalies of parallelism. The authors derive the performance bounds of parallel OR-tree searches with respect to the best-first, depth-first, and breadth-first strategies, and verify these bounds by simulation. They show that a near-linear speedup can be achieved with respect to a large number of processors for parallel OR-tree searches. Using the bounds developed, the authors derive sufficient conditions for assuring that parallelism will not degrade performance and necessary conditions for allowing parallelism to have a speedup greater than the ratio of the numbers of processors. These bounds and conditions provide the theoretical foundation for determining the number of processors required to assure a near-linear speedup.

Li, Guo-Jie↗

Parallel computation in a three-dimensional elastic-plastic finite-element analysis

A CRAY parallel processing technique called autotasking was implemented in a three-dimensional elasto-plastic finite-element code. The technique was evaluated on two CRAY supercomputers, a CRAY 2 and a CRAY Y-MP. Autotasking was implemented in all major portions of the code, except the matrix equations solver. Compiler directives alone were not able to properly multitask the code; user-inserted directives were required to achieve better performance. It was noted that the connect time, rather than wall-clock time, was more appropriate to determine speedup in multiuser environments. For a typical example problem, a speedup of 2.1 (1.8 when the solution time was included) was achieved in a dedicated environment and 1.7 (1.6 with solution time) in a multiuser environment on a four-processor CRAY 2 supercomputer. The speedup on a three-processor CRAY Y-MP was about 2.4 (2.0 with solution time) in a multiuser environment.

Shivakumar, K. N.↗

Preconditioned conjugate gradient methods for the Navier-Stokes equations

A preconditioned Krylov subspace method (GMRES) is used to solve the linear systems of equations formed at each time-integration step of the unsteady, two-dimensional, compressible Navier-Stokes equations of fluid flow. The Navier-Stokes equations are cast in an implicit, upwind finite-volume, flux-split formulation. Several preconditioning techniques are investigated to enhance the efficiency and convergence rate of the implicit solver based on the GMRES algorithm. The superiority of the new solver is established by comparisons with a conventional implicit solver, namely line Gauss-Seidel relaxation (LGSR). Computational test results for low-speed (incompressible flow over a backward-facing step at Mach 0.1), transonic flow (trailing edge flow in a transonic turbine cascade), and hypersonic flow (shock-on-shock interactions on a cylindrical leading edge at Mach 6.0) are presented. For the Mach 0.1 case, overall speedup factors of up to 17 (in terms of time-steps) and 15 (in terms of CPU time on a CRAY-YMP/8) are found in favor of the preconditioned GMRES solver, when compared with the LGSR solver. The corresponding speedup factors for the transonic flow case are 17 and 23, respectively. The hypersonic flow case shows slightly lower speedup factors of 9 and 13, respectively. The study of preconditioners conducted in this research reveals that a new LUSGS-type preconditioner is much more efficient than a conventional incomplete LU-type preconditioner.

Ajmani, Kumud↗

Scalability of Parallel Spatial Direct Numerical Simulations on Intel Hypercube and IBM SP1 and SP2

The implementation and performance of a parallel spatial direct numerical simulation (PSDNS) approach on the Intel iPSC/860 hypercube and IBM SP1 and SP2 parallel computers is documented. Spatially evolving disturbances associated with the laminar-to-turbulent transition in boundary-layer flows are computed with the PSDNS code. The feasibility of using the PSDNS to perform transition studies on these computers is examined. The results indicate that PSDNS approach can effectively be parallelized on a distributed-memory parallel machine by remapping the distributed data structure during the course of the calculation. Scalability information is provided to estimate computational costs to match the actual costs relative to changes in the number of grid points. By increasing the number of processors, slower than linear speedups are achieved with optimized (machine-dependent library) routines. This slower than linear speedup results because the computational cost is dominated by FFT routine, which yields less than ideal speedups. By using appropriate compile options and optimized library routines on the SP1, the serial code achieves 52-56 M ops on a single node of the SP1 (45 percent of theoretical peak performance). The actual performance of the PSDNS code on the SP1 is evaluated with a "real world" simulation that consists of 1.7 million grid points. One time step of this simulation is calculated on eight nodes of the SP1 in the same time as required by a Cray Y/MP supercomputer. For the same simulation, 32-nodes of the SP1 and SP2 are required to reach the performance of a Cray C-90. A 32 node SP1 (SP2) configuration is 2.9 (4.6) times faster than a Cray Y/MP for this simulation, while the hypercube is roughly 2 times slower than the Y/MP for this application. KEY WORDS: Spatial direct numerical simulations; incompressible viscous flows; spectral methods; finite differences; parallel computing.

Joslin, Ronald D.↗

The Optimization of Trained and Untrained Image Classification Algorithms for Use on Large Spatial Datasets

The HARVIST project seeks to automatically provide an accurate, interactive interface to predict crop yield over the entire United States. In order to accomplish this goal, large images must be quickly and automatically classified by crop type. Current trained and untrained classification algorithms, while accurate, are highly inefficient when operating on large datasets. This project sought to develop new variants of two standard trained and untrained classification algorithms that are optimized to take advantage of the spatial nature of image data. The first algorithm, harvist-cluster, utilizes divide-and-conquer techniques to precluster an image in the hopes of increasing overall clustering speed. The second algorithm, harvistSVM, utilizes support vector machines (SVMs), a type of trained classifier. It seeks to increase classification speed by applying a "meta-SVM" to a quick (but inaccurate) SVM to approximate a slower, yet more accurate, SVM. Speedups were achieved by tuning the algorithm to quickly identify when the quick SVM was incorrect, and then reclassifying low-confidence pixels as necessary. Comparing the classification speeds of both algorithms to known baselines showed a slight speedup for large values of k (the number of clusters) for harvist-cluster, and a significant speedup for harvistSVM. Future work aims to automate the parameter tuning process required for harvistSVM, and further improve classification accuracy and speed. Additionally, this research will move documents created in Canvas into ArcGIS. The launch of the Mars Reconnaissance Orbiter (MRO) will provide a wealth of image data such as global maps of Martian weather and high resolution global images of Mars. The ability to store this new data in a georeferenced format will support future Mars missions by providing data for landing site selection and the search for water on Mars.

Kocurek, Michael J.↗

Parallel Computation of the Jacobian Matrix for Nonlinear Equation Solvers Using MATLAB

Demonstrating speedup for parallel code on a multicore shared memory PC can be challenging in MATLAB due to underlying parallel operations that are often opaque to the user. This can limit potential for improvement of serial code even for the so-called embarrassingly parallel applications. One such application is the computation of the Jacobian matrix inherent to most nonlinear equation solvers. Computation of this matrix represents the primary bottleneck in nonlinear solver speed such that commercial finite element (FE) and multi-body-dynamic (MBD) codes attempt to minimize computations. A timing study using MATLAB's Parallel Computing Toolbox was performed for numerical computation of the Jacobian. Several approaches for implementing parallel code were investigated while only the single program multiple data (spmd) method using composite objects provided positive results. Parallel code speedup is demonstrated but the goal of linear speedup through the addition of processors was not achieved due to PC architecture.

Rose, Geoffrey K.↗

SCORPIO: A Scalable Two-Phase Parallel I/O Library With Application To A Large Scale Subsurface Simulator

Inefficient parallel I/O is known to be a major bottleneck among scientific applications employed on supercomputers as the number of processor cores grows into the thousands. Our prior experience indicated that parallel I/O libraries such as HDF5 that rely on MPI-IO do not scale well beyond 10K processor cores, especially on parallel file systems (like Lustre) with single point of resource contention. Our previous optimization efforts for a massively parallel multi-phase and multi-component subsurface simulator (PFLOTRAN) led to a two-phase I/O approach at the application level where a set of designated processes participate in the I/O process by splitting the I/O operation into a communication phase and a disk I/O phase. The designated I/O processes are created by splitting the MPI global communicator into multiple sub-communicators. The root process in each sub-communicator is responsible for performing the I/O operations for the entire group and then distributing the data to rest of the group. This approach resulted in over 25X speedup in HDF I/O read performance and 3X speedup in write performance for PFLOTRAN at over 100K processor cores on the ORNL Jaguar supercomputer. This research describes the design and development of a general purpose parallel I/O library, SCORPIO (SCalable block-ORiented Parallel I/O) that incorporates our optimized two-phase I/O approach. The library provides a simplified higher level abstraction to the user, sitting atop existing parallel I/O libraries (such as HDF5) and implements optimized I/O access patterns that can scale on larger number of processors. Performance results with standard benchmark problems and PFLOTRAN indicate that our library is able to maintain the same speedups as before with the added flexibility of being applicable to a wider range of I/O intensive applications.

High Performance Computing↗

Limits on the Efficiency of Event-Based Algorithms for Monte Carlo Neutron Transport

The traditional form of parallelism in Monte Carlo particle transport simulations, wherein each individual particle history is considered a unit of work, does not lend itself well to data-level parallelism. Event-based algorithms, which were originally used for simulations on vector processors, may offer a path toward better utilizing data-level parallelism in modern computer architectures. In this study, a simple model is developed for estimating the efficiency of the event-based particle transport algorithm under two sets of assumptions. Data collected from simulations of four reactor problems using OpenMC was then used in conjunction with the models to calculate the speedup due to vectorization as a function of two parameters: the size of the particle bank and the vector width. When each event type is assumed to have constant execution time, the achievable speedup is directly related to the particle bank size. We observed that the bank size generally needs to be at least 20 times greater than vector size in order to achieve vector efficiency greater than 90%. When the execution times for events are allowed to vary, however, the vector speedup is also limited by differences in execution time for events being carried out in a single event-iteration. For some problems, this implies that vector effciencies over 50% may not be attainable. While there are many factors impacting performance of an event-based algorithm that are not captured by our model, it nevertheless provides insights into factors that may be limiting in a real implementation.

Romano, Paul K.↗

m-CUBES An efficient and portable implementation of multi-dimensional integration for gpus

The task of multi-dimensional numerical integration is frequently encountered in physics and other scientific fields, e.g., in modeling the effects of systematic uncertainties in physical systems and in Bayesian parameter estimation. Multi-dimensional integration is often time-prohibitive on CPUs. Efficient implementation on many-core architectures is challenging as the workload across the integration space cannot be predicted a priori. We propose m-Cubes, a novel implementation of the well-known Vegas algorithm for execution on GPUs. Vegas transforms integration variables followed by calculation of a Monte Carlo integral estimate using adaptive partitioning of the resulting space. m-Cubes improves performance on GPUs by maintaining relatively uniform workload across the processors. As a result, our optimized Cuda implementation for Nvidia GPUs outperforms parallelization approaches proposed in past literature. We further demonstrate the efficiency of m-Cubes by evaluating a six-dimensional integral from a cosmology application, achieving significant speedup and greater precision than the CUBA library's CPU implementation of VEGAS. We also evaluate m-Cubes on a standard integrand test suite. m-Cubes outperforms the serial implementations of the Cuba and GSL libraries by orders of magnitude speedup while maintaining comparable accuracy. Our approach yields a speedup of at least 10 when compared against publicly available Monte Carlo based GPU implementations. In summary, m-Cubes can solve integrals that are prohibitively expensive using standard libraries and custom implementations. A modern C++ interface header-only implementation makes m-Cubes portable, allowing its utilization in complicated pipelines with easy to define stateful integrals. Compatibility with non-Nvidia GPUs is achieved with our initial implementation of m-Cubes using the Kokkos framework.

Sakiotis, Ioannis↗

Generalizing mkFit and its Application to HL-LHC

mkFit is an implementation of the Kalman filter-based track reconstruction algorithm that exploits both thread- and data-level parallelism. In the past few years the project transitioned from the R&D phase to deployment in the Run-3 offline workflow of the CMS experiment. The CMS tracking performs a series of iterations, targeting reconstruction of tracks of increasing difficulty after removing hits associated to tracks found in previous iterations. mkFit has been adopted for several of the tracking iterations, which contribute to the majority of reconstructed tracks. When tested in the standard conditions for production jobs, speedups in track pattern recognition are on average of the order of 3.5x for the iterations where it is used (3-7x depending on the iteration). Multiple factors contribute to the observed speedups, including vectorization and a lightweight geometry description, as well as improved memory management and single precision. Efficient vectorization is achieved with both the icc and the gcc (default in CMSSW) compilers and relies on a dedicated library for small matrix operations, Matriplex, which has recently been released in a public repository. While the mkFit geometry description already featured levels of abstraction from the actual Phase-1 CMS tracker, several components of the implementations were still tied to that specific geometry. We have further generalized the geometry description and the configuration of the run-time parameters, in order to enable support for the Phase-2 upgraded tracker geometry for the HL-LHC and potentially other detector configurations. The implementation strategy and high-level code changes required for the HL-LHC geometry are presented. Speedups in track building from mkFit imply that track fitting becomes a comparably time consuming step of the tracking chain.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Compute-Efficient Deep Learning: Algorithmic Trends and Opportunities

Although deep learning has made great progress in recent years, the exploding economic and environmental costs of training neural networks are becoming unsustainable. To address this problem, there has been a great deal of research on algorithmically-efficient deep learning, which seeks to reduce training costs not at the hardware or implementation level, but through changes in the semantics of the training program. In this paper, we present a structured and comprehensive overview of the research in this field. First, we formalize the algorithmic speedup problem, then we use fundamental building blocks of algorithmically efficient training to develop a taxonomy. Our taxonomy highlights commonalities of seemingly disparate methods and reveals current research gaps. Next, we present evaluation best practices to enable comprehensive, fair, and reliable comparisons of speedup techniques. To further aid research and applications, we discuss common bottlenecks in the training pipeline (illustrated via experiments) and offer taxonomic mitigation strategies for them. Finally, we highlight some unsolved research challenges and present promising future directions.

97 MATHEMATICS AND COMPUTING↗