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 1,099 records · Page 61

Real-time optimization of multi-cell industrial evaporative cooling towers using machine learning and particle swarm optimization

Existing electrical generating stations must operate with greater flexibility due to increasing renewable energy penetration on the electrical grid, and many coal-fired power stations have transitioned away from baseload operation to load-following operation to aid in grid stability. In cases where multiple independently controlled cooling tower cells are used in parallel for the cooling purposes of such stations, there is an opportunity to increase plant efficiency through data-driven optimization across their full load ranges. This work presents a novel application of real-time optimization using machine learning and particle swarm optimization on a multi-cell induced-draft cooling tower servicing a coal-fired power station under variable load. This is the first work to demonstrate simultaneous optimization of a multi-cell cooling tower, in addition to using machine learning for closed-loop control on a cooling tower. A novel control configuration is presented that ensures original control logic is not adversely affected and that the overall plant process is not disrupted using only existing hardware and operational data. To verify this methodology, the 12 independent cooling tower cells are simulated in parallel using historic operating data to demonstrate the effectiveness of real-time optimization compared to current practice. An artificial neural network is trained to predict overall cooling tower power consumption using only operational data and ambient conditions with an R2 value of greater than 0.96. The real-time optimization using particle swarm yields 6.7% annual energy usage savings compared to current practices, although the extent of the real-time savings varies greatly with both plant load and environmental conditions. This is particularly significant for a variable load situation because frequent ramping typically results in reduced overall efficiency. Furthermore, this proposed AI-based solution presents an opportunity to improve the overall heat rate of a load-following coal-fired power plant without the need to perform extensive first-principles modeling or add additional hardware to the cooling tower, resulting in more resources conserved and less overall emissions per unit of electricity generated.

42 ENGINEERING↗

Scalable parallel Newton-Krylov solvers for discontinuous Galerkin discretizations

We present techniques for implicit solution of discontinuous Galerkin discretizations of the Navier-Stokes equations on parallel computers. While a block-Jacobi method is simple and straight-forward to parallelize, its convergence properties are poor except for simple problems. Therefore, we consider Newton-GMRES methods preconditioned with block-incomplete LU factorizations, with optimized element orderings based on a minimum discarded fill (MDF) approach. We discuss the difficulties with the parallelization of these methods, but also show that with a simple domain decomposition approach, most of the advantages of the block-ILU over the block-Jacobi preconditioner are still retained. The convergence is further improved by incorporating the matrix connectivities into the mesh partitioning process, which aims at minimizing the errors introduced from separating the partitions. We demonstrate the performance of the schemes for realistic two- and three-dimensional flow problems.

97 MATHEMATICS AND COMPUTING↗

Collaborating CPU and GPU for large-scale high-order CFD simulations with complex grids on the TianHe-1A supercomputer

Programming and optimizing complex, real-world CFD codes on current many-core accelerated HPC systems is very challenging, especially when collaborating CPUs and accelerators to fully tap the potential of heterogeneous systems. In this paper, with a tri-level hybrid and heterogeneous programming model using MPI + OpenMP + CUDA, we port and optimize our high-order multi-block structured CFD software HOSTA on the GPU-accelerated TianHe-1A supercomputer. HOSTA adopts two self-developed high-order compact definite difference schemes WCNS and HDCS that can simulate flows with complex geometries. We present a dual-level parallelization scheme for efficient multi-block computation on GPUs and perform particular kernel optimizations for high-order CFD schemes. The GPU-only approach achieves a speedup of about 1.3 when comparing one Tesla M2050 GPU with two Xeon X5670 CPUs. To achieve a greater speedup, we collaborate CPU and GPU for HOSTA instead of using a naive GPU-only approach. We present a novel scheme to balance the loads between the store-poor GPU and the store-rich CPU. Taking CPU and GPU load balance into account, we improve the maximum simulation problem size per TianHe-1A node for HOSTA by 2.3×, meanwhile the collaborative approach can improve the performance by around 45% compared to the GPU-only approach. Further, to scale HOSTA on TianHe-1A, we propose a gather/scatter optimization to minimize PCI-e data transfer times for ghost and singularity data of 3D grid blocks, and overlap the collaborative computation and communication as far as possible using some advanced CUDA and MPI features. Scalability tests show that HOSTA can achieve a parallel efficiency of above 60% on 1024 TianHe-1A nodes. With our method, we have successfully simulated an EET high-lift airfoil configuration containing 800M cells and China's large civil airplane configuration containing 150M cells. To our best knowledge, those are the largest-scale CPU–GPU collaborative simulations that solve realistic CFD problems with both complex configurations and high-order schemes.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Hands-Free Transcranial Color Doppler Probe

Current transcranial color Doppler (TCD) transducer probes are bulky and difficult to move in tiny increments to search and optimize TCD signals. This invention provides miniature motions of a TCD transducer probe to optimize TCD signals. The mechanical probe uses a spherical bearing in guiding and locating the tilting crystal face. The lateral motion of the crystal face as it tilts across the full range of motion was achieved by minimizing the distance between the pivot location and the crystal face. The smallest commonly available metal spherical bearing was used with an outer diameter of 12 mm, a 3-mm tall retaining ring, and 5-mm overall height. Small geared motors were used that would provide sufficient power in a very compact package. After confirming the validity of the basic positioning concept, optimization design loops were completed to yield the final design. A parallel motor configuration was used to minimize the amount of space wasted inside the probe case while minimizing the overall case dimensions. The distance from the front edge of the crystal to the edge of the case was also minimized to allow positioning of the probe very close to the ear on the temporal lobe. The mechanical probe is able to achieve a +/-20deg tip and tilt with smooth repeatable action in a very compact package. The enclosed probe is about 7 cm long, 4 cm wide, and 1.8 cm tall. The device is compact, hands-free, and can be adjusted via an innovative touchscreen. Positioning of the probe to the head is performed via conventional transducer gels and pillows. This device is amendable to having advanced software, which could intelligently focus and optimize the TCD signal.

Chin, Robert↗

Evolutionary Computational Methods for Identifying Emergent Behavior in Autonomous Systems

A technique based on Evolutionary Computational Methods (ECMs) was developed that allows for the automated optimization of complex computationally modeled systems, such as autonomous systems. The primary technology, which enables the ECM to find optimal solutions in complex search spaces, derives from evolutionary algorithms such as the genetic algorithm and differential evolution. These methods are based on biological processes, particularly genetics, and define an iterative process that evolves parameter sets into an optimum. Evolutionary computation is a method that operates on a population of existing computational-based engineering models (or simulators) and competes them using biologically inspired genetic operators on large parallel cluster computers. The result is the ability to automatically find design optimizations and trades, and thereby greatly amplify the role of the system engineer.

Terrile, Richard J.↗

Structural design using equilibrium programming formulations

Solutions to increasingly larger structural optimization problems are desired. However, computational resources are strained to meet this need. New methods will be required to solve increasingly larger problems. The present approaches to solving large-scale problems involve approximations for the constraints of structural optimization problems and/or decomposition of the problem into multiple subproblems that can be solved in parallel. An area of game theory, equilibrium programming (also known as noncooperative game theory), can be used to unify these existing approaches from a theoretical point of view (considering the existence and optimality of solutions), and be used as a framework for the development of new methods for solving large-scale optimization problems. Equilibrium programming theory is described, and existing design techniques such as fully stressed design and constraint approximations are shown to fit within its framework. Two new structural design formulations are also derived. The first new formulation is another approximation technique which is a general updating scheme for the sensitivity derivatives of design constraints. The second new formulation uses a substructure-based decomposition of the structure for analysis and sensitivity calculations. Significant computational benefits of the new formulations compared with a conventional method are demonstrated.

Scotti, Stephen J.↗

Towards Native Execution of Deep Learning on a Leadership-Class HPC System

Large parallel machines generally offer the best parallel performance with "native execution" that is achieved using codes developed with the optimized compilers, communication libraries, and runtimes offered on the machines. In this paper, we report and analyze performance results from native execution of deep learning on a leadership-class high-performance computing (HPC) system. Using our new code called DeepEx, we present a study of the parallel speed up and convergence rates of learning achieved with native parallel execution. In the trade-off between computational parallelism and synchronized convergence, we first focus on maximizing parallelism while still obtaining convergence. Scaling results are reported from execution on up to 15,000 GPUs using two scientific data sets from atom microscopy and protein folding applications, and also using the popular ImageNet data set. In terms of the traditional measure of parallel speed up, excellent scaling is observed up to 12,000 GPUs. Additionally, accounting for convergence rates of deep learning accuracy or error, a deep learning-specific metric called "learning speed up" is also tracked. The performance results indicate the need to evaluate parallel deep learning execution in terms of learning speed up, and point to additional directions for improved exploitation of high-end HPC systems.

Yoginath, Srikanth↗

SWAP-Assembler 2: Optimization of De Novo Genome Assembler at Large Scale

In this paper, we analyze and optimize the most time-consuming steps of the SWAP-Assembler, a parallel genome assembler, so that it can scale to a large number of cores for huge genomes with the size of sequencing data ranging from terabyes to petabytes. According to the performance analysis results, the most time-consuming steps are input parallelization, k-mer graph construction, and graph simplification (edge merging). For the input parallelization, the input data is divided into virtual fragments with nearly equal size, and the start position and end position of each fragment are automatically separated at the beginning of the reads. In k-mer graph construction, in order to improve the communication efficiency, the message size is kept constant between any two processes by proportionally increasing the number of nucleotides to the number of processes in the input parallelization step for each round. The memory usage is also decreased because only a small part of the input data is processed in each round. With graph simplification, the communication protocol reduces the number of communication loops from four to two loops and decreases the idle communication time. The optimized assembler is denoted as SWAP-Assembler 2 (SWAP2). In our experiments using a 1000 Genomes project dataset of 4 terabytes (the largest dataset ever used for assembling) on the supercomputer Mira, the results show that SWAP2 scales to 131,072 cores with an efficiency of 40%. We also compared our work with both the HipMER assembler and the SWAP-Assembler. On the Yanhuang dataset of 300 gigabytes, SWAP2 shows a 3X speedup and 4X better scalability compared with the HipMer assembler and is 45 times faster than the SWAP-Assembler. The SWAP2 software is available at https://sourceforge.net/projects/swapassembler.

Meng, Jintao↗

Optimizing Grain Boundary Structures with LAMMPS Using Evolutionary Algorithms

Grain boundary structure optimization is an important part of materials modeling. Current methods for grain boundary structure optimization involve inefficient, time-consuming processes that do not fully explore the interface parameter space. Evolutionary algorithms have recently been demonstrated to be effective at determining both stable and metastable grain boundary interface structures. In this work, we demonstrate the use of GBOpt, a grain boundary structure optimization software designed to use the Large-scale Atomic/Molecular Massively Parallel Simulation (LAMMPS) software to efficiently determine grain boundary structures. We demonstrate that a only a few manipulations, namely atom insertion, atom removal, and relative grain displacement, are sufficient to explore much of the grain boundary structure parameter space. The efficacy of this approach is demonstrated on an FCC Ni system, and a BCC Fe system. The computational cost is compared against the gamma-surface sampling approach to demonstrate performance improvement.

Evolutionary algorithms↗

An Atmospheric General Circulation Model with Chemistry for the CRAY T3E: Design, Performance Optimization and Coupling to an Ocean Model

The design, implementation and performance optimization on the CRAY T3E of an atmospheric general circulation model (AGCM) which includes the transport of, and chemical reactions among, an arbitrary number of constituents is reviewed. The parallel implementation is based on a two-dimensional (longitude and latitude) data domain decomposition. Initial optimization efforts centered on minimizing the impact of substantial static and weakly-dynamic load imbalances among processors through load redistribution schemes. Recent optimization efforts have centered on single-node optimization. Strategies employed include loop unrolling, both manually and through the compiler, the use of an optimized assembler-code library for special function calls, and restructuring of parts of the code to improve data locality. Data exchanges and synchronizations involved in coupling different data-distributed models can account for a significant fraction of the running time. Therefore, the required scattering and gathering of data must be optimized. In systems such as the T3E, there is much more aggregate bandwidth in the total system than in any particular processor. This suggests a distributed design. The design and implementation of a such distributed 'Data Broker' as a means to efficiently couple the components of our climate system model is described.

Farrara, John D.↗

Integrated Sensing and Controls for Coal Gasification - Development of Model-Based Controls for GE's Gasifier and Syngas Cooler (Final Technical Report)

This report summarizes the achievements and final results of this program. The objective of this program is to develop a comprehensive systems approach to integrated design of sensing and control systems for an Integrated Gasification Combined Cycle (IGCC) plant, using advanced model-based techniques. In particular, this program is focused on the model-based sensing and control system design for the core gasification section of an IGCC plant. The overall approach consists of (i) developing a first-principles physics-based dynamic model of the gasification section, (ii) performing model-reduction where needed to derive low-order models suitable for controls analysis and design, (iii) developing a sensing system solution combining online sensors with model-based estimation for important process variables not measured directly, and (iv) optimizing the steady-state and transient operation of the plant for normal operation as well as for startup using model predictive controls (MPC). Initially, available process unit models were implemented in a common platform using Matlab/Simulink{reg_sign}, and appropriate model reduction and model updates were performed to obtain the overall gasification section dynamic model. Also, a set of sensor packages were developed through extensive lab testing and implemented in the Tampa Electric Company IGCC plant at Polk power station in 2009, to measure temperature and strain in the radiant syngas cooler (RSC). Plant operation data was also used to validate the overall gasification section model. The overall dynamic model was then used to develop a sensing solution including a set of online sensors coupled with model-based estimation using nonlinear extended Kalman filter (EKF). Its performance in terms of estimating key unmeasured variables like gasifier temperature, carbon conversion, etc., was studied through extensive simulations in the presence sensing errors (noise and bias) and modeling errors (e.g. unknown gasifier kinetics, RSC fouling). In parallel, an MPC solution was initially developed using ideal sensing to optimize the plant operation during startup pre-heating as well as steady state and transient operation under normal high-pressure conditions, e.g. part-load, base-load, load transition and fuel changes. The MPC simulation studies showed significant improvements both for startup pre-heating and for normal operation. Finally, the EKF and MPC solutions were coupled to achieve the integrated sensing and control solution and its performance was studied through extensive steady state and transient simulations in the presence of sensor and modeling errors. The results of each task in the program and overall conclusions are summarized in this final report.

01 COAL, LIGNITE, AND PEAT↗

Optimization of Collimator Trajectory in Volumetric Modulated Arc Therapy: Development and Evaluation for Paraspinal SBRT

Purpose: To develop a collimator trajectory optimization paradigm for volumetric modulated arc therapy (VMAT) and evaluate this technique in paraspinal stereotactic body radiation therapy (SBRT). Method and Materials: We propose a novel VMAT paradigm, Coll-VMAT, which integrates collimator rotation with synchronized gantry rotation, multileaf collimator (MLC) motion, and dose-rate modulation. At each gantry angle a principal component analysis (PCA) is applied to calculate the primary cord orientation. The collimator angle is then aligned so that MLC travel is parallel to the PCA-derived direction. An in-house VMAT optimization follows the geometry-based collimator trajectory optimization to obtain the optimal MLC position and monitor units (MU) at each gantry angle. A treatment planning study of five paraspinal SBRT patients compared Coll-VMAT to standard VMAT (fixed collimator angle) and static field IMRT plans. Plan evaluation statistics included planning target volume (PTV) V95%, PTV-D95%, cord-D05%, and total beam-on time. Results: Variation of collimator angle in Coll-VMAT plans ranges from 26 deg. to 54 deg., with a median of 40 deg. Patient-averaged PTV V95% (94.6% Coll-VMAT vs. 92.1% VMAT and 93.3% IMRT) and D95% (22.5 Gy vs. 21.4 Gy and 22.0 Gy, respectively) are highest with Coll-VMAT, and cord D05% (9.8 Gy vs. 10.0 Gy and 11.7 Gy) is lowest. Total beam-on time with Coll-VMAT (5,164 MU) is comparable to standard VMAT (4,868 MU) and substantially lower than IMRT (13,283 MU). Conclusion: Collimator trajectory optimization-based VMAT provides an additional degree of freedom that can improve target coverage and cord sparing of paraspinal SBRT plans compared with standard VMAT and IMRT approaches.

07 ISOTOPE AND RADIATION SOURCES↗

Optimal control system design of an acid gas removal unit for an IGCC power plants with CO2 capture

Future IGCC plants with CO{sub 2} capture should be operated optimally in the face of disturbances without violating operational and environmental constraints. To achieve this goal, a systematic approach is taken in this work to design the control system of a selective, dual-stage Selexol-based acid gas removal (AGR) unit for a commercial-scale integrated gasification combined cycle (IGCC) power plant with pre-combustion CO{sub 2} capture. The control system design is performed in two stages with the objective of minimizing the auxiliary power while satisfying operational and environmental constraints in the presence of measured and unmeasured disturbances. In the first stage of the control system design, a top-down analysis is used to analyze degrees of freedom, define an operational objective, identify important disturbances and operational/environmental constraints, and select the control variables. With the degrees of freedom, the process is optimized with relation to the operational objective at nominal operation as well as under the disturbances identified. Operational and environmental constraints active at all operations are chosen as control variables. From the results of the optimization studies, self-optimizing control variables are identified for further examination. Several methods are explored in this work for the selection of these self-optimizing control variables. Modifications made to the existing methods will be discussed in this presentation. Due to the very large number of candidate sets available for control variables and due to the complexity of the underlying optimization problem, solution of this problem is computationally expensive. For reducing the computation time, parallel computing is performed using the Distributed Computing Server (DCS®) and the Parallel Computing® toolbox from Mathworks®. The second stage is a bottom-up design of the control layers used for the operation of the process. First, the regulatory control layer is designed followed by the supervisory control layer. Finally, an optimization layer is designed. In this paper, the proposed two-stage control system design approach is applied to the AGR unit for an IGCC power plant with CO{sub 2} capture. Aspen Plus Dynamics® is used to develop the dynamic AGR process model while MATLAB is used to perform the control system design and for implementation of model predictive control (MPC).

97 MATHEMATICS AND COMPUTING↗

Non-Intrusive Parallel-in-Time Solvers for Partial Differential Equations (Final Report)

Many time-dependent problems and simulations are often modeled using Partial Differential Equations. Traditional modeling approaches that use sequential time-stepping are reaching a bottleneck in optimizing efficiency. The Center of Applied Science and Computing at Lawrence Livermore National Laboratory extensively works on parallelizing these algorithms to leverage the increasing computational power from the growing number of processors in computer hardware. In particular, they aim to design non-intrusive algorithms that can generalize to a variety of problems and sizes without requiring additional information from or modifications on the original problems. Multigrid Reduction in Time (MGRIT) is a parallel-in-time algorithm that is designed to be non-intrusive. This project focuses on increasing the efficiency of MGRIT by approximating the coarse-grid operator using machine learning approaches as a means to find the most non-intrusive, or general, solution.

97 MATHEMATICS AND COMPUTING↗

Monte Carlo-based treatment planning system calculation engine for microbeam radiation therapy

Purpose: Microbeam radiation therapy (MRT) is a synchrotron radiotherapy technique that explores the limits of the dose-volume effect. Preclinical studies have shown that MRT irradiations (arrays of 25-75-{mu}m-wide microbeams spaced by 200-400 {mu}m) are able to eradicate highly aggressive animal tumor models while healthy tissue is preserved. These promising results have provided the basis for the forthcoming clinical trials at the ID17 Biomedical Beamline of the European Synchrotron Radiation Facility (ESRF). The first step includes irradiation of pets (cats and dogs) as a milestone before treatment of human patients. Within this context, accurate dose calculations are required. The distinct features of both beam generation and irradiation geometry in MRT with respect to conventional techniques require the development of a specific MRT treatment planning system (TPS). In particular, a Monte Carlo (MC)-based calculation engine for the MRT TPS has been developed in this work. Experimental verification in heterogeneous phantoms and optimization of the computation time have also been performed. Methods: The penelope/penEasy MC code was used to compute dose distributions from a realistic beam source model. Experimental verification was carried out by means of radiochromic films placed within heterogeneous slab-phantoms. Once validation was completed, dose computations in a virtual model of a patient, reconstructed from computed tomography (CT) images, were performed. To this end, decoupling of the CT image voxel grid (a few cubic millimeter volume) to the dose bin grid, which has micrometer dimensions in the transversal direction of the microbeams, was performed. Optimization of the simulation parameters, the use of variance-reduction (VR) techniques, and other methods, such as the parallelization of the simulations, were applied in order to speed up the dose computation. Results: Good agreement between MC simulations and experimental results was achieved, even at the interfaces between two different media. Optimization of the simulation parameters and the use of VR techniques saved a significant amount of computation time. Finally, parallelization of the simulations improved even further the calculation time, which reached 1 day for a typical irradiation case envisaged in the forthcoming clinical trials in MRT. An example of MRT treatment in a dog's head is presented, showing the performance of the calculation engine. Conclusions: The development of the first MC-based calculation engine for the future TPS devoted to MRT has been accomplished. This will constitute an essential tool for the future clinical trials on pets at the ESRF. The MC engine is able to calculate dose distributions in micrometer-sized bins in complex voxelized CT structures in a reasonable amount of time. Minimization of the computation time by using several approaches has led to timings that are adequate for pet radiotherapy at synchrotron facilities. The next step will consist in its integration into a user-friendly graphical front-end.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Computational Approaches to Simulation and Optimization of Global Aircraft Trajectories

This study examines three possible approaches to improving the speed in generating wind-optimal routes for air traffic at the national or global level. They are: (a) using the resources of a supercomputer, (b) running the computations on multiple commercially available computers and (c) implementing those same algorithms into NASAs Future ATM Concepts Evaluation Tool (FACET) and compares those to a standard implementation run on a single CPU. Wind-optimal aircraft trajectories are computed using global air traffic schedules. The run time and wait time on the supercomputer for trajectory optimization using various numbers of CPUs ranging from 80 to 10,240 units are compared with the total computational time for running the same computation on a single desktop computer and on multiple commercially available computers for potential computational enhancement through parallel processing on the computer clusters. This study also re-implements the trajectory optimization algorithm for further reduction of computational time through algorithm modifications and integrates that with FACET to facilitate the use of the new features which calculate time-optimal routes between worldwide airport pairs in a wind field for use with existing FACET applications. The implementations of trajectory optimization algorithms use MATLAB, Python, and Java programming languages. The performance evaluations are done by comparing their computational efficiencies and based on the potential application of optimized trajectories. The paper shows that in the absence of special privileges on a supercomputer, a cluster of commercially available computers provides a feasible approach for national and global air traffic system studies.

global air traffic optimization↗

Computational Approaches to Simulation and Optimization of Global Aircraft Trajectories

This study examines three possible approaches to improving the speed in generating wind-optimal routes for air traffic at the national or global level. They are: (a) using the resources of a supercomputer, (b) running the computations on multiple commercially available computers and (c) implementing those same algorithms into NASA’s Future ATM Concepts Evaluation Tool (FACET) and compares those to a standard implementation run on a single CPU. Wind-optimal aircraft trajectories are computed using global air traffic schedules. The run time and wait time on the supercomputer for trajectory optimization using various numbers of CPUs ranging from 80 to 10,240 units are compared with the total computational time for running the same computation on a single desktop computer and on multiple commercially available computers for potential computational enhancement through parallel processing on the computer clusters. This study also re-implements the trajectory optimization algorithm for further reduction of computational time through algorithm modifications and integrates that with FACET to facilitate the use of the new features which calculate time-optimal routes between worldwide airport pairs in a wind field for use with existing FACET applications. The implementations of trajectory optimization algorithms use MATLAB, Python, and Java programming languages. The performance evaluations are done by comparing their computational efficiencies and based on the potential application of optimized trajectories. The paper shows that in the absence of special privileges on a supercomputer, a cluster of commercially available computers provides a good option for computing wind-optimal trajectories for national and global air traffic system studies.

Ng, Hok K.↗

Optimization of the hop-byte metric for effective topology aware mapping

Suitable mapping of processes to the nodes of a massively parallel machine can substantially improve communication performance by reducing network congestion. The hop-byte metric has been used as a measure of the quality of such a mapping by several recent works. Optimizing this metric is NP hard, and thus heuristics are applied. However, the heuristics proposed so far do not directly try to optimize this metric. Rather, they use some intuitive methods for reducing congestion and use the metric just to evaluate the quality of the mapping. In fact, heuristics intending to optimize other metrics too don't directly optimize for them, but, rather, use the metric to evaluate the results of the heuristic. In contrast, we pose the mapping problem with the hop-byte metric as a quadratic assignment problem and use a heuristic to directly optimize for this metric. We evaluate our approach on realistic node allocations obtained on the Kraken system at NICS. Our approach yields values for the metric that are up to 75% lower than the default mapping and 66 % lower than existing heuristics. However, the time taken to produce the mapping can be substantially more, which makes this suitable for somewhat static, though possibly irregular, communication patterns. We introduce new heuristics that reduce the time taken to be comparable to that of existing fast heuristics, while still producing mappings of higher quality than existing ones. We also use theoretical lower bounds to suggest that our mapping may be close to optimal, at least for medium sized problems. Consequently, our work can also provide insight into the tradeoff between mapping quality and time taken by other mapping heuristics.

Sudheer, C. D.↗