Search NASA⌕ Search

SEARCH · Search NASA

Results for “parallel computing algorithms”

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 1,423 records · Page 79

An efficient MPI/OpenMP parallelization of the Hartree–Fock–Roothaan method for the first generation of Intel® Xeon Phi™ processor architecture

The Hartree-Fock (HF) method in the quantum chemistry package GAMESS represents one of the most irregular algorithms in computation today. Major steps in the calculation are the irregular computation of electron repulsion integrals (ERIs) and the building of the Fock matrix. These are the central components of the main Self Consistent Field (SCF) loop, the key hotspot in Electronic Structure (ES) codes. By threading the MPI ranks in the official release of the GAMESS code, we not only speed up the main SCF loop (4x to 6x for large systems), but also achieve a significant (>2x) reduction in the overall memory footprint. These improvements are a direct consequence of memory access optimizations within the MPI ranks. We benchmark our implementation against the official release of the GAMESS code on the Intel R Xeon PhiTM supercomputer. Here, scaling numbers are reported on up to 7,680 cores on Intel Xeon Phi coprocessors.

97 MATHEMATICS AND COMPUTING↗

Efficient Delaunay Tessellation through K-D Tree Decomposition

Delaunay tessellations are fundamental data structures in computational geometry. They are important in data analysis, where they can represent the geometry of a point set or approximate its density. The algorithms for computing these tessellations at scale perform poorly when the input data is unbalanced. We investigate the use of k-d trees to evenly distribute points among processes and compare two strategies for picking split points between domain regions. Because resulting point distributions no longer satisfy the assumptions of existing parallel Delaunay algorithms, we develop a new parallel algorithm that adapts to its input and prove its correctness. We evaluate the new algorithm using two late-stage cosmology datasets. The new running times are up to 50 times faster using k-d tree compared with regular grid decomposition. Moreover, in the unbalanced data sets, decomposing the domain into a k-d tree is up to five times faster than decomposing it into a regular grid.

Morozov, Dmitriy↗

CHOLLA: A NEW MASSIVELY PARALLEL HYDRODYNAMICS CODE FOR ASTROPHYSICAL SIMULATION

We present Computational Hydrodynamics On ParaLLel Architectures (Cholla ), a new three-dimensional hydrodynamics code that harnesses the power of graphics processing units (GPUs) to accelerate astrophysical simulations. Cholla models the Euler equations on a static mesh using state-of-the-art techniques, including the unsplit Corner Transport Upwind algorithm, a variety of exact and approximate Riemann solvers, and multiple spatial reconstruction techniques including the piecewise parabolic method (PPM). Using GPUs, Cholla evolves the fluid properties of thousands of cells simultaneously and can update over 10 million cells per GPU-second while using an exact Riemann solver and PPM reconstruction. Owing to the massively parallel architecture of GPUs and the design of the Cholla code, astrophysical simulations with physically interesting grid resolutions (≳256{sup 3}) can easily be computed on a single device. We use the Message Passing Interface library to extend calculations onto multiple devices and demonstrate nearly ideal scaling beyond 64 GPUs. A suite of test problems highlights the physical accuracy of our modeling and provides a useful comparison to other codes. We then use Cholla to simulate the interaction of a shock wave with a gas cloud in the interstellar medium, showing that the evolution of the cloud is highly dependent on its density structure. We reconcile the computed mixing time of a turbulent cloud with a realistic density distribution destroyed by a strong shock with the existing analytic theory for spherical cloud destruction by describing the system in terms of its median gas density.

79 ASTRONOMY AND ASTROPHYSICS↗

Scientific data analysis on data-parallel platforms.

As scientific computing users migrate to petaflop platforms that promise to generate multi-terabyte datasets, there is a growing need in the community to be able to embed sophisticated analysis algorithms in the computing platforms' storage systems. Data Warehouse Appliances (DWAs) are attractive for this work, due to their ability to store and process massive datasets efficiently. While DWAs have been utilized effectively in data-mining and informatics applications, they remain largely unproven in scientific workloads. In this paper we present our experiences in adapting two mesh analysis algorithms to function on five different DWA architectures: two Netezza database appliances, an XtremeData dbX database, a LexisNexis DAS, and multiple Hadoop MapReduce clusters. The main contribution of this work is insight into the differences between these DWAs from a user's perspective. In addition, we present performance measurements for ten DWA systems to help understand the impact of different architectural trade-offs in these systems.

99 GENERAL AND MISCELLANEOUS↗

TurboRVB: A many-body toolkit for ab initio electronic simulations by quantum Monte Carlo

TurboRVB is a computational package for ab initio Quantum Monte Carlo (QMC) simulations of both molecular and bulk electronic systems. The code implements two types of well established QMC algorithms: Variational Monte Carlo (VMC) and diffusion Monte Carlo in its robust and efficient lattice regularized variant. A key feature of the code is the possibility of using strongly correlated many-body wave functions (WFs), capable of describing several materials with very high accuracy, even when standard mean-field approaches [e.g., density functional theory (DFT)] fail. The electronic WF is obtained by applying a Jastrow factor, which takes into account dynamical correlations, to the most general mean-field ground state, written either as an antisymmetrized geminal power with spin-singlet pairing or as a Pfaffian, including both singlet and triplet correlations. This WF can be viewed as an efficient implementation of the so-called resonating valence bond (RVB) Ansatz, first proposed by Pauling and Anderson in quantum chemistry [L. Pauling, The Nature of the Chemical Bond (Cornell University Press, 1960)] and condensed matter physics [P.W. Anderson, Mat. Res. Bull 8, 153 (1973)], respectively. The RVB Ansatz implemented in TurboRVB has a large variational freedom, including the Jastrow correlated Slater determinant as its simplest, but nontrivial case. Moreover, it has the remarkable advantage of remaining with an affordable computational cost, proportional to the one spent for the evaluation of a single Slater determinant. Therefore, its application to large systems is computationally feasible. The WF is expanded in a localized basis set. Several basis set functions are implemented, such as Gaussian, Slater, and mixed types, with no restriction on the choice of their contraction. The code implements the adjoint algorithmic differentiation that enables a very efficient evaluation of energy derivatives, comprising the ionic forces. Thus, one can perform structural optimizations and molecular dynamics in the canonical NVT ensemble at the VMC level. For the electronic part, a full WF optimization (Jastrow and antisymmetric parts together) is made possible, thanks to state-of-the-art stochastic algorithms for energy minimization. In the optimization procedure, the first guess can be obtained at the mean-field level by a built-in DFT driver. The code was efficiently parallelized by using a hybrid MPI-OpenMP protocol, which is also an ideal environment for exploiting the computational power of modern Graphics Processing Unit accelerators.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

An efficient parallel simulation of unsteady blood flows in patient-specific pulmonary artery

Simulation of blood flows in the pulmonary artery provides some insight into certain diseases by examining the relationship between some continuum metrics, eg, the wall shear stress acting on the vascular endothelium, which responds to flow-induced mechanical forces by releasing vasodilators/constrictors. V. Kheyfets, in his previous work, studies numerically a patient-specific pulmonary circulation to show that decreasing wall shear stress is correlated with increasing pulmonary vascular impedance. We develop a scalable parallel algorithm based on domain decomposition methods to investigate an unsteady model with patient-specific pulsatile waveforms as the inlet boundary condition. The unsteady model offers tremendously more information about the dynamic behavior of the flow field, but computationally speaking, the simulation is a lot more expensive since a problem which is similar to the steady-state problem has to be solved many times, and therefore, the traditional sequential approach is not suitable anymore. We show computationally that simulations using the proposed parallel approach with up to 10 000 processor cores can be obtained with much reduced compute time. Finally, this makes the technology potentially usable for the routine study of the dynamic behavior of blood flows in the pulmonary artery, in particular, the changes of the blood flows and the wall shear stress in the spatial and temporal dimensions.

60 APPLIED LIFE SCIENCES↗

Approximate algorithms for partitioning and assignment problems

The problem of optimally assigning the modules of a parallel/pipelined program over the processors of a multiple computer system under certain restrictions on the interconnection structure of the program as well as the multiple computer system was considered. For a variety of such programs it is possible to find linear time if a partition of the program exists in which the load on any processor is within a certain bound. This method, when combined with a binary search over a finite range, provides an approximate solution to the partitioning problem. The specific problems considered were: a chain structured parallel program over a chain-like computer system, multiple chain-like programs over a host-satellite system, and a tree structured parallel program over a host-satellite system. For a problem with m modules and n processors, the complexity of the algorithm is no worse than O(mnlog(W sub T/epsilon)), where W sub T is the cost of assigning all modules to one processor and epsilon the desired accuracy.

Iqbal, M. A.↗

A Scalable Gaussian Process Analysis Algorithm for Biomass Monitoring

Biomass monitoring is vital for studying the carbon cycle of earth's ecosystem and has several significant implications, especially in the context of understanding climate change and its impacts. Recently, several change detection methods have been proposed to identify land cover changes in temporal profiles (time series) of vegetation collected using remote sensing instruments, but do not satisfy one or both of the two requirements of the biomass monitoring problem, i.e., {\em operating in online mode} and {\em handling periodic time series}. In this paper, we adapt Gaussian process regression to detect changes in such time series in an online fashion. While Gaussian process (GP) have been widely used as a kernel based learning method for regression and classification, their applicability to massive spatio-temporal data sets, such as remote sensing data, has been limited owing to the high computational costs involved. We focus on addressing the scalability issues associated with the proposed GP based change detection algorithm. This paper makes several significant contributions. First, we propose a GP based online time series change detection algorithm and demonstrate its effectiveness in detecting different types of changes in {\em Normalized Difference Vegetation Index} (NDVI) data obtained from a study area in Iowa, USA. Second, we propose an efficient Toeplitz matrix based solution which significantly improves the computational complexity and memory requirements of the proposed GP based method. Specifically, the proposed solution can analyze a time series of length $t$ in $O(t^2)$ time while maintaining a $O(t)$ memory footprint, compared to the $O(t^3)$ time and $O(t^2)$ memory requirement of standard matrix manipulation based methods. Third, we describe a parallel version of the proposed solution which can be used to simultaneously analyze a large number of time series. We study three different parallel implementations: using threads, MPI, and a hybrid implementation using threads and MPI. Experimental results show that the hybrid implementation scales better than the multi-threaded and MPI based implementations. The application of the proposed scalable algorithm is demonstrated in analyzing massive remote sensing observation data. The hybrid implementation, using 1536 computing cores, can analyze an NDVI data set for the Iowa study area in nearly 5 seconds, while a serial algorithm, using standard Cholesky decomposition based routines, takes several days to process the same data set.

09 BIOMASS FUELS↗

Scaling and Quality of Modularity Optimization Methods for Graph Clustering

Real-world graphs exhibit structures known as “communities” or “clusters” consisting of a group of vertices with relatively high connectivity between them, as compared to the rest of the vertices in the network. Graph clustering or community detection is a fundamental graph operation used to analyze real-world graphs occurring in the areas of computational biology, cybersecurity, electrical grids, etc. Similar to other graph algorithms, owing to irregular memory accesses and inherently sequential nature, current algorithms for community detection are challenging to parallelize. However, in order to analyze large networks, it is important to develop scalable parallel implementations of graph clustering that are capable of exploiting the architectural features of modern supercomputers. In response to the 2019 Streaming Graph Challenge, we present quality and performance analysis of our distributed- memory community detection using Vite, which is our distributed memory implementation of the popular Louvain method, on the ALCF Theta supercomputer. Clustering methods such as Louvain that rely on modularity maximization are known to suffer from the resolution limit problem, preventing identification of clusters of certain sizes. Hence, we also include quality analysis of our shared-memory implementation of the Fast-tracking Resistance method, in comparison with Louvain on the challenge datasets. Furthermore, we introduce an edge-balanced graph distribution for our distributed memory implementation, that significantly reduces communication, offering up to 80% improvement in the overall execution time. In addition to performance/quality analysis, we also include details on the power/energy consumption, and memory traffic of the distributed-memory clustering implementation using real-world graphs with over a billion edges.

Ghosh, Sayan↗

Performance Models for the Spike Banded Linear System Solver

With availability of large-scale parallel platforms comprised of tens-of-thousands of processors and beyond, there is significant impetus for the development of scalable parallel sparse linear system solvers and preconditioners. An integral part of this design process is the development of performance models capable of predicting performance and providing accurate cost models for the solvers and preconditioners. There has been some work in the past on characterizing performance of the iterative solvers themselves. In this paper, we investigate the problem of characterizing performance and scalability of banded preconditioners. Recent work has demonstrated the superior convergence properties and robustness of banded preconditioners, compared to state-of-the-art ILU family of preconditioners as well as algebraic multigrid preconditioners. Furthermore, when used in conjunction with efficient banded solvers, banded preconditioners are capable of significantly faster time-to-solution. Our banded solver, the Truncated Spike algorithm is specifically designed for parallel performance and tolerance to deep memory hierarchies. Its regular structure is also highly amenable to accurate performance characterization. Using these characteristics, we derive the following results in this paper: (i) we develop parallel formulations of the Truncated Spike solver, (ii) we develop a highly accurate pseudo-analytical parallel performance model for our solver, (iii) we show excellent predication capabilities of our model – based on which we argue the high scalability of our solver. Our pseudo-analytical performance model is based on analytical performance characterization of each phase of our solver. These analytical models are then parameterized using actual runtime information on target platforms. An important consequence of our performance models is that they reveal underlying performance bottlenecks in both serial and parallel formulations. All of our results are validated on diverse heterogeneous multiclusters – platforms for which performance prediction is particularly challenging. Finally, we provide predict the scalability of the Spike algorithm using up to 65,536 cores with our model. In this paper we extend the results presented in the Ninth International Symposium on Parallel and Distributed Computing.

Manguoglu, Murat↗

High-speed computerized tomography

The development of a high-speed reconstruction processor and a channelized architecture to use with a high-resolution tomographic unit is discussed with attention to the convolution reconstruction algorithm. By means of this algorithm, input data and intermediate result precision required throughout the algorithm execution have been studied with computer simulation using profile data derived from mathematically simulated test objects and experimental animal data. A prototype section for a highly parallel all-digital system executes 60 million arithmetic operations per second, and the full-scale version is expected to reconstruct 500 to 1000 cross sections per second.

Swartzlander, E. E., Jr.↗

Mathematical model partitioning and packing for parallel computer calculation

This paper deals with the development of multiprocessor simulations from a serial set of ordinary differential equations describing a physical system. The identification of computational parallelism within the model equations is discussed. A technique is presented for identifying this parallelism and for partitioning the equations for parallel solution on a multiprocessor. Next, an algorithm which packs the equations into a minimum number of processors is described. The results of applying the packing algorithm to a turboshaft engine model are presented.

Arpasi, Dale J.↗

Efficiency of group implicit concurrent algorithms for transient finite element analysis

The performance of group implicit algorithms is assessed on actual concurrent computers. It is shown that, as the number of subdomains is increased, performance enhancements are derived from two sources: the increased parallelism in the computations; and a reduction in equation solving effort. Moreover, these two performance enhancements are synergistic, in the sense that the corresponding speed-ups are multiplied, rather than merely added. Simulations on a 32-node hypercube are presented for which the interprocessor communications efficiencies obtained are consistently in excess of 90 percent.

Ortiz, M.↗

Ordered fast fourier transforms on a massively parallel hypercube multiprocessor

Design alternatives for ordered Fast Fourier Transformation (FFT) algorithms were examined on massively parallel hypercube multiprocessors such as the Connection Machine. Particular emphasis is placed on reducing communication which is known to dominate the overall computing time. To this end, the order and computational phases of the FFT were combined, and the sequence to processor maps that reduce communication were used. The class of ordered transforms is expanded to include any FFT in which the order of the transform is the same as that of the input sequence. Two such orderings are examined, namely, standard-order and A-order which can be implemented with equal ease on the Connection Machine where orderings are determined by geometries and priorities. If the sequence has N = 2 exp r elements and the hypercube has P = 2 exp d processors, then a standard-order FFT can be implemented with d + r/2 + 1 parallel transmissions. An A-order sequence can be transformed with 2d - r/2 parallel transmissions which is r - d + 1 fewer than the standard order. A parallel method for computing the trigonometric coefficients is presented that does not use trigonometric functions or interprocessor communication. A performance of 0.9 GFLOPS was obtained for an A-order transform on the Connection Machine.

Tong, Charles↗

A data distributed parallel algorithm for ray-traced volume rendering

This paper presents a divide-and-conquer ray-traced volume rendering algorithm and a parallel image compositing method, along with their implementation and performance on the Connection Machine CM-5, and networked workstations. This algorithm distributes both the data and the computations to individual processing units to achieve fast, high-quality rendering of high-resolution data. The volume data, once distributed, is left intact. The processing nodes perform local ray tracing of their subvolume concurrently. No communication between processing units is needed during this locally ray-tracing process. A subimage is generated by each processing unit and the final image is obtained by compositing subimages in the proper order, which can be determined a priori. Test results on both the CM-5 and a group of networked workstations demonstrate the practicality of our rendering algorithm and compositing method.

Ma, Kwan-Liu↗

An Efficient GPU-Accelerated Multi-Source Global Fit Pipeline for LISA Data Analysis

The large-scale analysis task of deciphering gravitational wave signals in the LISA data stream will be difficult, requiring a large amount of computational resources and extensive development of computational methods. Its high dimensionality, multiple model types, and complicated noise profile require a global fit to all parameters and input models simultaneously. In this work, we detail our global fit algorithm, called “Erebor,” designed to accomplish this challenging task. It is capable of analysing current state-of-the-art datasets and then growing into the future as more pieces of the pipeline are completed and added. We describe our pipeline strategy, the algorithmic setup, and the results from our analysis of the LDC2A Sangria dataset, which contains Massive Black Hole Binaries, compact Galactic Binaries, and a parameterized noise spectrum whose parameters are unknown to the user. The Erebor algorithm includes three unique and very useful contributions: GPU acceleration for enhanced computational efficiency; ensemble MCMC sampling with multiple MCMC walkers per temperature for better mixing and parallelized sample creation; and special online updates to reversible-jump (or trans-dimensional) sampling distributions to ensure sampler mixing and accurate initial estimates for detectable sources in the data. We recover posterior distributions for all 15 (6) of the injected MBHBs in the LDC2A training (hidden) dataset. We catalog ∼12000 Galactic Binaries (∼8000 as high confidence detections) for both the training and hidden datasets. All of the sources and their posterior distributions are provided in publicly available catalogs.

LISA global fit↗

Portability: A Necessary Approach for Future Scientific Software

Today's world of scientific software for High Energy Physics (HEP) is powered by x86 code, while the future will be much more reliant on accelerators like GPUs and FPGAs. The portable parallelization strategies (PPS) project of the High Energy Physics Center for Computational Excellence (HEP/CCE) is investigating solutions for portability techniques that will allow the coding of an algorithm once, and the ability to execute it on a variety of hardware products from many vendors, especially including accelerators. We think without these solutions, the scientific success of our experiments and endeavors is in danger, as software development could be expert driven and costly to be able to run on available hardware infrastructure. We think the best solution for the community would be an extension to the C++ standard with a very low entry bar for users, supporting all hardware forms and vendors. We are very far from that ideal though. We argue that in the future, as a community, we need to request and work on portability solutions and strive to reach this ideal.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

HydraGNN_Predictive_GFM_2024 - Ensemble of predictive graph foundation models for ground state atomistic materials modeling

We provide the ensemble of fifteen pre-trained graph foundation models (GFMs) for atomistic materials modeling applications. Each one of the fifteen GFMs has been trained on five open-source datasets that (once aggregated) amount to over 154 million atomistic structures, which cover over two-thirds of the natural elements of the periodic table and that comprises a broad set of organic and inorganic compounds. This vast set of atomistic structures comprises ground state configurations that are dynamically stable (i.e., equilibrated structures with atomic forces approximately close to zero values) as well as dynamically unstable structures (i.e., non-equilibrium structures with non-negligible non-zero values of atomic forces). The ensemble of datasets aggregated does NOT include excited states. The datasets have been curated to remove atomistic structures with spectral norm of the force tensor above 100 eV/angstrom. Moreover, a linear term of the energy was computed for each dataset using a linear regression model that uses the chemical concentration of each natural element as regressor. The linear term predicted by the linear regression model has been subtracted from each original energy value to perform a re-alignment of the energy values across different electronic structures approximation theories performed to generate the diverse multi-source, multi-fidelity datasets. The folder "ADIOS_files" contains the set of pre-processed datasets in Adaptable I/O System (ADIOS) format (https://www.exascaleproject.org/research-project/adios/) that have been used for the development and training of GFMs in this work. The "ADIOS_files" directory contains 6 sub-directories named as follows: - ANI1x-v3.bp - MPTrj-v3.bp - OC2020-20M-v3.bp - OC2020-v3.bp - OC2022-v3.bp - qm7x-v3.bp Each sub-directory contains the pre-processed datasets converted in Adaptable I/O System (ADIOS) format (https://www.exascaleproject.org/research-project/adios/) that have been used to the development, training, and performance testing of the ensemble go predictive graph foundation models. Each GFM was developed using HydraGNN (https://github.com/ORNL/HydraGNN) as underlying graph neural network (GNN) architecture. The multi-task learning (MTL) capability of HydraGNN was used to simultaneously train the GFMs on labeled values for direct predictions of energy (a total system property of an atomistic structure that measures the chemical stability) and atomic forces (an atomic level property of an atomistic structure that measures the dynamical stability). The hyper parameters of the GFM have been tuned using scalable hyperparameter optimization (HPO) algorithms implemented in the software DeepHyper (https://github.com/deephyper/deephyper). The pre-training of each HPO trial was performed using distributed data parallelism (DDP) to scale the training across 128 compute nodes of the exascale OLCF supercomputer Frontier. Each HPO trial was trained only for 10 epochs and an early stopping was performed to avoid wasting significant computational resources on GNN architectures that were clearly underperforming. For each HPO trial, the 'omnistat' tool developed by (AMD Research - Advanced Micro Device) was used to measure the total energy consumption in kWh. The ensemble of GFMs was obtained by selecting the fifteen best performing HPO trials. Four models have been selected for their clear advantage in accuracy, and these are the GFMs with IDs 229, 156, 147, 260. Additional eleven models have been selected based on judicious balance between accuracy and energy consumption needed for training, and these are the GFMs with IDs 165, 78, 137, 1, 175, 171, 181, 67, 179, 167, 351. Each selected GFM of the ensemble was continued to cumulate a total of at most 30 epochs. In some cases, the total number of epochs actually performed was les than 30 due to two combined factors: (1) the size of the GFM (i.e., the number of model parameters to train) and (2) the total wall-clock time for which the computational resources could be allocated on OLCF-Frontier. The "Ensemble_of_models" directory contains 15 sub-directories named as follows: - gfm_0.229 - gfm_0.156 - gfm_0.147 - gfm_0.260 - gfm_0.165 - gfm_0.78 - gfm_0.137 - gfm_0.1 - gfm_0.175 - gfm_0.171 - gfm_0.181 - gfm_0.67 - gfm_0.179 - gfm_0.167 - gfm_0.351 Each one of these sub-directories refers to one of the fifteen HPO trials that have been selected to continue the pre-training with at most 30 epochs. With each sub-directory associated with a specific HPO trial, the following files can be found: - config.json: file for argument parsing to develop and train an HydraGNN architecture - gfm_0.ID_epoch_N.pk: file with model parameters for HPO ID trial after N epochs of training The ensemble of fifteen GFM architectures was used for (1) ensemble averaging to stabilize the predictions of energy and atomic forces after pre-training for post-processing analysis and (2) ensemble uncertainty quantification (UQ). The code used to develop, pre-train, and load the pre-trained models for post-processing analysis is available on the ORNL-GitHub at the following link: https://github.com/ORNL/HydraGNN/tree/Predictive_GFM_2024

36 MATERIALS SCIENCE↗