Search NASA⌕ Search

SEARCH · Search NASA

Results for “Parallel optimization”

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 1,513 records · Page 84

Poster - 19: Investigation of Electron Reference Dosimetry Based on Optimal Chamber Shift

An addendum/revision to AAPM TG-51 electron reference dosimetry is highly expected to meet the clinical requirement with the increasing usage of new ion chambers not covered in TG-51. A recent study, Med. Phys. 41, 111701, proposed a new fitting equation for the beam quality conversion factor k’{sub Q} to a wide spectrum of chambers. In the study, an optimal Effective Point of Measurement (EPOM) from Monte Carlo calculations was recommended and the fitting parameters to k’{sub Q} was based on it. We investigated the absolute dose obtained based on the optimal EPOM method and the original TG-51 method with k’{sub R50} determined differently. The results showed that using the Markus curve is a better choice than the well-guarded chamber fitting for an IBA PPC-05 parallel plate chamber if we need to strictly follow the AAPM TG-51 protocol. We also examined the usage of the new fitting equation with measurement performed at the physical EPOM, instead of the optimal EPOM. The former is more readily determined and more practical in clinics. Our study indicated that the k’{sub Q} fitting based on the optimal EPOM can be used to measurement at the physical EPOM with no significant clinical impact. The inclusion of Farmer chamber gradient correction P{sub gr} in k’{sub Q}, as in the mentioned study, asks for the precise positioning of chamber center at dref. It is not recommended in clinics to avoid over-correction for low electron energies, especially for an institute having matching Linacs implemented.

60 APPLIED LIFE SCIENCES↗

Hyper-Differential Sensitivity Analysis of Uncertain Parameters in PDE-Constrained Optimization

Many problems in engineering and sciences require the solution of large scale optimization constrained by partial differential equations (PDEs). Though PDE-constrained optimization is itself challenging, most applications pose additional complexity, namely, uncertain parameters in the PDEs. Uncertainty quantification (UQ) is necessary to characterize, prioritize, and study the influence of these uncertain parameters. Sensitivity analysis, a classical tool in UQ, is frequently used to study the sensitivity of a model to uncertain parameters. In this article, we introduce "hyper-differential sensitivity analysis" which considers the sensitivity of the solution of a PDE-constrained optimization problem to uncertain parameters. Our approach is a goal-oriented analysis which may be viewed as a tool to complement other UQ methods in the service of decision making and robust design. We formally define hyper-differential sensitivity indices and highlight their relationship to the existing optimization and sensitivity analysis literatures. Assuming the presence of low rank structure in the parameter space, computational efficiency is achieved by leveraging a generalized singular value decomposition in conjunction with a randomized solver which converts the computational bottleneck of the algorithm into an embarrassingly parallel loop. Two multi-physics examples, consisting of nonlinear steady state control and transient linear inversion, demonstrate efficient identification of the uncertain parameters which have the greatest influence on the optimal solution.

97 MATHEMATICS AND COMPUTING↗

Optimal design of microtube recuperators for an indirect supercritical carbon dioxide recompression closed Brayton cycle

This paper presents a baseline design and optimization approach developed in Aspen Custom Modeler (ACM) for microtube shell-and-tube exchangers (MSTEs) used for high- and low-temperature recuperation in a 10 MWe indirect supercritical carbon dioxide (sCO 2 ) recompression closed Brayton cycle (RCBC). The MSTE-type recuperators are designed using one-dimensional models with thermal-hydraulic correlations appropriate for sCO 2 and properties models that capture considerable nonlinear changes in CO 2 properties near the critical and pseudo-critical points. Using the successive quadratic programming (SQP) algorithm in ACM, optimal recuperator designs are obtained for either custom or industry-standard microtubes considering constraints based on current advanced manufacturing techniques. The three decision variables are the number of tubes, tube pitch-to-diameter ratio, and tube diameter. Five different objective functions based on different key design measures are considered: minimization of total heat transfer area, heat exchanger volume, metal weight, thermal residence time, and maximization of compactness. Sensitivities studies indicate the constraint on the maximum number of tubes per shell does affect the number of parallel heat exchanger trains but not the tube selection, total number of tubes, tube length and other key design measures in the final optimal design when considering industry-standard tubes. In this study, the optimally designed high- and low-temperature recuperators have 47,000 3/32 inch tubes and 63,000 1/16 inch tubes, respectively. In addition, sensitivities to the design temperature approach and maximum allowable pressure drop are studied, since these specifications significantly impact the optimal design of the recuperators as well as the thermal efficiency and the economic performance of the entire sCO 2 Brayton cycle.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Parallel interior-point solver for block-structured nonlinear programs on SIMD/GPU architectures

Here, we investigate how to port the standard interior-point method to new exascale architectures for block-structured nonlinear programs with state equations. Computationally, we decompose the interior-point algorithm into two successive operations: the evaluation of the derivatives and the solution of the associated Karush-Kuhn-Tucker (KKT) linear system. Our method accelerates both operations using two levels of parallelism. First, we distribute the computations on multiple processes using coarse parallelism. Second, each process uses SIMD/GPU accelerators locally to accelerate the operations using fine-grained parallelism. The KKT system is reduced by eliminating the inequalities and the state variables from the corresponding equations. We demonstrate our method's capability on the supercomputer Polaris, a testbed for the future exascale Aurora system. Each node is equipped with four GPUs, a setup amenable to our two-level approach. Our experiments on the stochastic optimal power flow problem show that the reduction method is 50x faster than the sparse linear solver HSL MA57 running in serial on the CPU, and 6x faster than Pardiso running in parallel on CPU on the same number of processes.

97 MATHEMATICS AND COMPUTING↗

Smart-Pixel Array Processors Based on Optimal Cellular Neural Networks for Space Sensor Applications

A smart-pixel cellular neural network (CNN) with hardware annealing capability, digitally programmable synaptic weights, and multisensor parallel interface has been under development for advanced space sensor applications. The smart-pixel CNN architecture is a programmable multi-dimensional array of optoelectronic neurons which are locally connected with their local neurons and associated active-pixel sensors. Integration of the neuroprocessor in each processor node of a scalable multiprocessor system offers orders-of-magnitude computing performance enhancements for on-board real-time intelligent multisensor processing and control tasks of advanced small satellites. The smart-pixel CNN operation theory, architecture, design and implementation, and system applications are investigated in detail. The VLSI (Very Large Scale Integration) implementation feasibility was illustrated by a prototype smart-pixel 5x5 neuroprocessor array chip of active dimensions 1380 micron x 746 micron in a 2-micron CMOS technology.

Fang, Wai-Chi↗

Parallel transport dynamics for mixed quantum states with applications to time-dependent density functional theory

Direct simulation of the von Neumann dynamics for a general (pure or mixed) quantum state can often be expensive. One prominent example is the real-time time-dependent density functional theory (rt-TDDFT), a widely used framework for the first principle description of many-electron dynamics in chemical and materials systems. Practical rt-TDDFT calculations often avoid the direct simulation of the von Neumann equation, and solve instead a set of Schrödinger equations, of which the dynamics is equivalent to that of the von Neumann equation. However, the time step size employed by the Schrödinger dynamics is often much smaller. Here, in order to improve the time step size and the overall efficiency of the simulation, we generalize a recent work of the parallel transport (PT) dynamics for simulating pure states [An, Lin, Multiscale Model. Simul. 18, 612, 2020] to general quantum states. The PT dynamics provides the optimal gauge choice, and can employ a time step size comparable to that of the von Neumann dynamics. Going beyond the linear and near adiabatic regime in previous studies, we find that the error of the PT dynamics can be bounded by certain commutators between Hamiltonians, density matrices, and their derived quantities. Such a commutator structure is not present in the Schrödinger dynamics. We demonstrate that the parallel transport-implicit midpoint (PT-IM) method is a suitable method for simulating the PT dynamics, especially when the spectral radius of the Hamiltonian is large. The commutator structure of the error bound, and numerical results for model rt-TDDFT calculations in both linear and nonlinear regimes, confirm the advantage of the PT dynamics.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

A Polyhedral Transformation Framework for Compiler Optimization (Final Project Report)

The project developed the polyhedral compiler transformation module PolyOpt/Fortran in the ROSE compiler framework. PolyOpt/Fortran performs automated transformation of affine loop nests within FORTRAN programs for enhanced data locality and parallel execution. A FORTAN version of the Polybench library was also developed by the project. A third development was a dynamic analysis approach to gauge vectorization potential within loops of programs; software (DDVec) for automated instrumentation and dynamic analysis of programs was developed.

97 MATHEMATICS AND COMPUTING↗

Applications Performance Under MPL and MPI on NAS IBM SP2

On July 5, 1994, an IBM Scalable POWER parallel System (IBM SP2) with 64 nodes, was installed at the Numerical Aerodynamic Simulation (NAS) Facility Each node of NAS IBM SP2 is a "wide node" consisting of a RISC 6000/590 workstation module with a clock of 66.5 MHz which can perform four floating point operations per clock with a peak performance of 266 Mflop/s. By the end of 1994, 64 nodes of IBM SP2 will be upgraded to 160 nodes with a peak performance of 42.5 Gflop/s. An overview of the IBM SP2 hardware is presented. The basic understanding of architectural details of RS 6000/590 will help application scientists the porting, optimizing, and tuning of codes from other machines such as the CRAY C90 and the Paragon to the NAS SP2. Optimization techniques such as quad-word loading, effective utilization of two floating point units, and data cache optimization of RS 6000/590 is illustrated, with examples giving performance gains at each optimization step. The conversion of codes using Intel's message passing library NX to codes using native Message Passing Library (MPL) and the Message Passing Interface (NMI) library available on the IBM SP2 is illustrated. In particular, we will present the performance of Fast Fourier Transform (FFT) kernel from NAS Parallel Benchmarks (NPB) under MPL and MPI. We have also optimized some of Fortran BLAS 2 and BLAS 3 routines, e.g., the optimized Fortran DAXPY runs at 175 Mflop/s and optimized Fortran DGEMM runs at 230 Mflop/s per node. The performance of the NPB (Class B) on the IBM SP2 is compared with the CRAY C90, Intel Paragon, TMC CM-5E, and the CRAY T3D.

Saini, Subhash↗

MOOSE ProbML: Parallelized probabilistic machine learning and uncertainty quantification for computational energy applications

Here, this paper presents the development and demonstration of massively parallel probabilistic machine learning (ML) and uncertainty quantification (UQ) capabilities within the Multiphysics Object-Oriented Simulation Environment (MOOSE), an open-source computational platform for parallel finite element and finite volume analyses. In addressing the computational expense and uncertainties inherent in complex multiphysics simulations, this paper integrates Gaussian process (GP) variants, active learning, Bayesian inverse UQ, adaptive forward UQ, Bayesian optimization, evolutionary optimization, and Markov chain Monte Carlo (MCMC) within MOOSE. It also elaborates on the interaction among key MOOSE systems — Sampler, MultiApp, Reporter, and Surrogate — in enabling these capabilities. The modularity offered by these systems enables development of a multitude of probabilistic ML and UQ algorithms in MOOSE. Example code demonstrations include parallel active learning and parallel Bayesian inference via active learning. The impact of these developments is illustrated through five applications relevant to computational energy applications: UQ of nuclear fuel fission product release, using parallel active learning Bayesian inference; very rare events analysis in nuclear microreactors using active learning; advanced manufacturing process modeling using multi-output GPs (MOGPs) and dimensionality reduction; fluid flow using deep GPs (DGPs); and tritium transport model parameter optimization for fusion energy, using batch Bayesian optimization. These capabilities are part of the MOOSE framework.

97 - MATHEMATICS AND COMPUTING↗

Parametric Multi-Level Tiling of Imperfectly Nested Loops

Tiling is a critical loop transformation for generating high-performance code on modern architectures. Efficient generation of multilevel tiled code is essential to exploit several levels of parallelism and/or to maximize data reuse in deep memory hierarchies. Tiled loops with parameterized tile sizes (not compile time constants) facilitate runtime feedback and dynamic optimizations used in iterative compilation and automatic tuning. The existing parametric multilevel tiling approach has focused on transformation for perfectly nested loops, where all assignment statements are contained inside the innermost loop of a loop nest. Previous solutions to tiling for imperfect loop nests are limited to the case where tile sizes are fixed. In this paper, we present an approach to parameterized multilevel tiling for imperfectly nested loops. Our tiling algorithm generates loops that iterate over full rectangular tiles that are amenable for potential compiler optimizations such as register tiling. Experimental results using a number of computational benchmarks demonstrate the effectiveness of our tiling approach.

97 MATHEMATICS AND COMPUTING↗

Utilizing waste heat in wastewater treatment plants for water desalination: Modeling and Multi-Objective optimization of a Multi-Effect desalination system using Decision Tree Regression and Pelican optimization algorithm

This paper examines the feasibility of using waste heat from wastewater treatment plants (WWTPs) for water desalination. A model was developed to utilize waste heat from the gensets at As Samra WWTP in Jordan, using real data and TRNSYS® software to calculate available waste heat. The desalination process was then modeled with ASPEN PLUS® software, focusing on multi-effect desalination (MED). Both series and parallel configurations for the MED system were compared. The study investigated the effects of system feeding flow rate, feeding pressure, and heat input on productivity, performance ratio, and recovery ratio. The study also introduces a novel optimization technique combining machine learning and modern optimization algorithms to maximize system productivity and performance. Initially, a decision tree regression (DTR) model is developed to establish relationships between key independent variables (flow rate, feed pressure, and heat input) and dependent variables (productivity, performance ratio, and recovery ratio). The Pelican Optimization Algorithm (POA) is then used to identify the optimal values of the independent variables for maximum productivity and performance. The results show that using a series configuration yields a system productivity of 3984.2 kg/hr, a performance ratio of 3.78, and a recovery ratio of 0.991 at a feed flow rate of 4000 kg/hr, feed pressure of 3 bars, and heat input of 719 kW. Optimal productivity (4421 kg/hr), performance ratio (3.81), and recovery ratio (0.851) are achieved at a feed flow rate of 5166 kg/hr, feed pressure of 3.2 bars, and heat input of 794 kW. In conclusion, the techno-economic assessment indicates a levelized cost of water of 1.63 USD/m 3 for parallel configurations and 1.65 USD/m 3 for series configurations, with a payback period of less than two years.

42 ENGINEERING↗

Scalable Graph Analytics and HPC Operational Enhancement: Parallel Computing and ML/DL Innovations

Parallel computing plays a pivotal role in the efficient processing of large-scale graphs. Complex network analysis stands as a capti- vating research frontier, holding promise across diverse scientific domains such as sociology, biology, online media, and recommenda- tion systems. In this era, Machine Learning (ML) and Deep Learning (DL) have emerged as indispensable tools, underpinning remarkable technological achievements. Within this dynamic landscape, my research revolves around advancing parallel algorithms tailored for large-scale graph operations. To achieve this, I harness the power of cutting-edge technologies including OpenMP, MPI, HIP, and CUDA, on the High-Performance Computing (HPC) platforms to unlock optimal performance. I also apply ML/DL techniques to HPC operational data, to streamline the monitoring and maintenance of supercomputers, alleviating the complexities associated with their upkeep and enhancing user support. My research echoes the syn- ergy between parallel computing, large-scale graph analysis, and ML/DL, improving computational efficiency and user experience.

Sattar, Naw Safrin↗

Study of treatment as an alternative WM pathway: examples of I-graphite and metallic components such as steam generators - 15212

EDF is working continuously to optimize and secure its waste management pathways, leading in some case to develop appropriate treatment solutions that could be implemented parallel to reference solutions, complementary or as alternative solutions to already existing paths. This approach is illustrated by the examples of steam generators replaced during the operation of French PWR power plants, and of irradiated graphite (stacks and sleeves), coming from the decommissioning of UNGG reactors, which have been shut down. Concerning graphite wastes, the near surface disposal facility intended for this waste does not exist yet, and is currently giving rise to the research of a suitable site, and an evaluation of its performances. While accompanying ANDRA (French national agency for nuclear waste management), in charge of the development of the facility, and while waiting for more data to be available, EDF is continuing the consolidation of graphite wastes radiological inventory, and is working on the development of an alternative management solution, based on a thermal treatment process. Depending on the nature of the treatment, and on the choice of operating parameters, this treatment may contribute to reduce graphite wastes volume and/or promote its acceptance in a disposal facility. Reducing the radiological inventory by decontamination is one of the means to achieve these objectives. Concerning replaced steam generators, nearly 200 of them will have been removed by 2030, and stored on site waiting for their treatment, which represents about 65 000 tons of metal. Initially intended to be cut and packed on NPPs' sites during their dismantling, EDF is studying the economic value of having the steam generators treated in a centralized facility, which allows reducing the volume of final wastes by optimizing cutting operations, while considering recycling after melting. However, this studied pathway implies a number of difficulties to be overcome. In particular, the shipment of such large components to a centralized facility may be complex. Moreover, French regulatory framework does not allow recycling outside nuclear industry, therefore limiting recycling opportunities to storage containers for radioactive wastes. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

IS THERE AN UNACCOUNTED FOR EXCESS IN THE EXTRAGALACTIC COSMIC RADIO BACKGROUND?

Analyses of the distribution of absolute brightness temperature over the radio sky have recently led to suggestions that there exists a substantial unexplained extragalactic radio background. Consequently, there have been numerous attempts to place constraints on plausible origins of this 'excess'. We suggest here that this expectation of a large extragalactic background, over and above that contributed by the sources observed in the surveys, is based on an extremely simple geometry adopted to model the Galactic emission and the procedure adopted in the estimation of the extragalactic contribution. In this paper, we derive the extragalactic radio background from wide-field radio images using a more realistic modeling of the Galactic emission and decompose the sky maps at 150, 408, and 1420 MHz into anisotropic Galactic and isotropic extragalactic components. The anisotropic Galactic component is assumed to arise from a highly flattened spheroid representing the thick disk, embedded in a spherical halo, both centered at the Galactic center, along with Galactic sources, filamentary structures, and Galactic loops and spurs. All components are constrained to be positive and the optimization scheme minimizes the sky area occupied by the complex filaments. We show that in contrast with simple modeling of Galactic emission as a plane parallel slab, the more realistic modeling yields estimates for the uniform extragalactic brightness that are consistent with expectations from known extragalactic radio source populations.

79 ASTRONOMY AND ASTROPHYSICS↗

Modeling Solute Thermokinetics in LiCl-KCl Molten Salt for Nuclear Waste Separation (NEUP 10-418)

Recovery of actinides is an integral part of a closed nuclear fuel cycle. Pyrometallurgical nuclear fuel recycling processes have been developed in the past for recovering actinides from spent metallic and nitride fuels. The process is essentially to dissolve the spent fuel in a molten salt and then extract just the actinides for reuse in a reactor. Extraction is typically done through electrorefining, which involves electrochemical reduction of the dissolved actinides and plating onto a cathode. Knowledge of a number of basic thermokinetic properties of salts and salt-fuel mixtures is necessary for optimizing present and developing new approaches for pyrometallurgical waste processing. The properties of salt-fuel mixtures are presently being studied, but there are so many solutes and varying concentrations that direct experimental investigation is prohibitively time consuming and expensive (particularly for radioactive elements like Pu). Therefore, there is a need to reduce the number of required experiments through modeling of salt and salt-fuel mixture properties. This project will develop first-principles-based molecular modeling and simulation approaches to predict fundamental thermokinetic properties of dissolved actinides and fission products in molten salts. The focus of the proposed work is on property changes with higher concentrations (up to 5 mol%) of dissolved fuel components, where there is still very limited experimental data. The properties predicted with the modeling will be density, which is used to assess the amount of dissolved material in the salt; diffusion coefficients, which can control rates of material transport during separation; and solute activity, which determines total solubility and reduction potentials used during electrorefining. The work will focus on La, Sr, and U, which are chosen to include the important distinct categories of lanthanides, alkali earths, and actinides, respectively. Studies will be performed using LiCl-KCl salt at the eutectic composition (58 mol% LiCl, 42 mol% KCl), which is used for treating spent EBR-II fuel. The same process being used for EBRII fuel is currently being studied for widespread international implementation. The methods will focus on first-principles and first- principles derived interatomic potential based simulations, primarily using molecular dynamics. Results will be validated against existing literature and parallel ongoing experimental efforts. The simulation results will be of value for interpreting experimental results, validating analytical models, and for optimizing waste separation by potentially developing new salt configurations and operating conditions.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Parallel Adjective High-Order CFD Simulations Characterizing SOFIA Cavity Acoustics

This paper presents large-scale MPI-parallel computational uid dynamics simulations for the Stratospheric Observatory for Infrared Astronomy (SOFIA). SOFIA is an airborne, 2.5-meter infrared telescope mounted in an open cavity in the aft fuselage of a Boeing 747SP. These simulations focus on how the unsteady ow eld inside and over the cavity interferes with the optical path and mounting structure of the telescope. A temporally fourth-order accurate Runge-Kutta, and spatially fth-order accurate WENO- 5Z scheme was used to perform implicit large eddy simulations. An immersed boundary method provides automated gridding for complex geometries and natural coupling to a block-structured Cartesian adaptive mesh re nement framework. Strong scaling studies using NASA's Pleiades supercomputer with up to 32k CPU cores and 4 billion compu- tational cells shows excellent scaling. Dynamic load balancing based on execution time on individual AMR blocks addresses irregular numerical cost associated with blocks con- taining boundaries. Limits to scaling beyond 32k cores are identi ed, and targeted code optimizations are discussed.

CFD↗

Parallel beamlet dose calculation via beamlet contexts in a distributed multi‐GPU framework

Purpose Dose calculation is one of the most computationally intensive, yet essential tasks in the treatment planning process. With the recent interest in automatic beam orientation and arc trajectory optimization techniques, there is a great need for more efficient model‐based dose calculation algorithms that can accommodate hundreds to thousands of beam candidates at once. Foundational work has shown the translation of dose calculation algorithms to graphical processing units ( GPU s), lending to remarkable gains in processing efficiency. But these methods provide parallelization of dose for only a single beamlet, serializing the calculation of multiple beamlets and under‐utilizing the potential of modern GPU s. In this paper, the authors propose a framework enabling parallel computation of many beamlet doses using a novel beamlet context transformation and further embed this approach in a scalable network of multi‐ GPU computational nodes. Methods The proposed context‐based transformation separates beamlet‐local density and TERMA into distinct beamlet contexts that independently provide sufficient data for beamlet dose calculation. Beamlet contexts are arranged in a composite context array with dosimetric isolation, and the context array is subjected to a GPU collapsed‐cone convolution superposition procedure, producing the set of beamlet‐specific dose distributions in a single pass. Dose from each context is converted to a sparse representation for efficient storage and retrieval during treatment plan optimization. The context radius is a new parameter permitting flexibility between the speed and fidelity of the dose calculation process. A distributed manager‐worker architecture is constructed around the context‐based GPU dose calculation approach supporting an arbitrary number of worker nodes and resident GPU s. Phantom experiments were executed to verify the accuracy of the context‐based approach compared to Monte Carlo and a reference CPU ‐ CCCS implementation for single beamlets and broad beams composed by addition of beamlets. Dose for representative 4 π beam sets was calculated in lung and prostate cases to compare its efficiency with that of an existing beamlet‐sequential GPU ‐ CCCS implementation. Code profiling was also performed to evaluate the scalability of the framework across many networked GPU s. Results The dosimetric accuracy of the context‐based method displays <1.35% and 2.35% average error from the existing serialized CPU ‐ CCCS algorithm and Monte Carlo simulation for beamlet‐specific PDD s in water and slab phantoms, respectively. The context‐based method demonstrates substantial speedup of up to two orders of magnitude over the beamlet‐sequential GPU ‐ CCCS method in the tested configurations. The context‐based framework demonstrates near linear scaling in the number of distributed compute nodes and GPU s employed, indicating that it is flexible enough to meet the performance requirements of most users by simply increasing the hardware utilization. Conclusions The context‐based approach demonstrates a new expectation of performance for beamlet‐based dose calculation methods. This approach has been successful in accelerating the dose calculation process for very large‐scale treatment planning problems ‐ such as automatic 4 π IMRT beam orientation and VMAT arc trajectory selection, with hundreds of thousands of beamlets ‐ in clinically feasible timeframes. The flexibility of this framework makes it as a strong candidate for use in a variety of other very large‐scale treatment planning tasks and clinical workflows.

Neph, Ryan↗

A Flexible Forwarding Scheme to Improve Latency-Bound Irregular P2P Communication in MPI

We propose an algorithm to efficiently perform latency-bound communication scenarios that consist of many small messages. In these parallel scenarios, processes typically pass around a lot of small-sized messages of a few KBs of size. Performing communication operations with P2P MPI routines or collective MPI routines (including neighborhood collectives) in such scenarios may not always yield the optimal results and may not resolve the latency bottleneck. To this end, we develop a regular structure called virtual process topology (VPT) on which the messages can be communicated in a structured and controlled manner. Using parameters of this topology, one can tune the rate of aggression in tackling the latency costs. We demonstrate that our communication algorithm is preferable to MPI P2P and collective routines for latency-bound communication and it can easily be adapted only by replacing calls to MPI routines in a parallel application. We show how to adapt existing topology-aware mapping heuristics to address the volume overhead due to communicating messages on the VPT. Moreover, we propose a novel swap-based mapping heuristic to address this overhead by optimizing the maximum volume handled by a process. Experiments on synthetic communication graphs as well as real-world applications such as parallel Canonical Polyadic sparse tensor decomposition and parallel sparse matrix-dense matrix multiplication show that our approach is a powerful way of overcoming the bottlenecks posed by sparse and latency-bound irregular communication.

communication algorithm↗