Search NASA⌕ Search

SEARCH · Search NASA

Results for “code sharing”

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 163 records · Page 9
All sourcesNASA NTRSDOE OSTI
DOE OSTI2016

AiiDA: automated interactive infrastructure and database for computational science

Computational science has seen in the last decades a spectacular rise in the scope, breadth, and depth of its efforts. Notwithstanding this prevalence and impact, it is often still performed using the renaissance model of individual artisans gathered in a workshop, under the guidance of an established practitioner. Great benefits could follow instead from adopting concepts and tools coming from computer science to manage, preserve, and share these computational efforts. We illustrate here our paradigm sustaining such vision, based around the four pillars of Automation, Data, Environment, and Sharing. We then discuss its implementation in the open-source AiiDA platform (http://www.aiida.net), that has been tuned first to the demands of computational materials science. AiiDA’s design is based on directed acyclic graphs to track the provenance of data and calculations, and ensure preservation and searchability. Remote computational resources are managed transparently, and automation is coupled with data storage to ensure reproducibility. Last, complex sequences of calculations can be encoded into scientific workflows. We believe that AiiDA’s design and its sharing capabilities will encourage the creation of social ecosystems to disseminate codes, data, and scientific workflows.

Materials Science↗
DOE OSTI2016

An Efficient Multicore Implementation of a Novel HSS-Structured Multifrontal Solver Using Randomized Sampling

Here, we present a sparse linear system solver that is based on a multifrontal variant of Gaussian elimination and exploits low-rank approximation of the resulting dense frontal matrices. We use hierarchically semiseparable (HSS) matrices, which have low-rank off-diagonal blocks, to approximate the frontal matrices. For HSS matrix construction, a randomized sampling algorithm is used together with interpolative decompositions. The combination of the randomized compression with a fast ULV HSS factoriz ation leads to a solver with lower computational complexity than the standard multifrontal method for many applications, resulting in speedups up to 7 fold for problems in our test suite. The implementation targets many-core systems by using task parallelism with dynamic runtime scheduling. Numerical experiments show performance improvements over state-of-the-art sparse direct solvers. The implementation achieves high performance and good scalability on a range of modern shared memory parallel systems, including the Intel Xeon Phi (MIC). The code is part of a software package called STRUMPACK - STRUctured Matrices PACKage, which also has a distributed memory component for dense rank-structured matrices.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Efforts to enhance reproducibility in a human performance research project

Background: Ensuring the validity of results from funded programs is a critical concern for agencies that sponsor biological research. In recent years, the open science movement has sought to promote reproducibility by encouraging sharing not only of finished manuscripts but also of data and code supporting their findings. While these innovations have lent support to third-party efforts to replicate calculations underlying key results in the scientific literature, fields of inquiry where privacy considerations or other sensitivities preclude the broad distribution of raw data or analysis may require a more targeted approach to promote the quality of research output. Methods: We describe efforts oriented toward this goal that were implemented in one human performance research program, Measuring Biological Aptitude, organized by the Defense Advanced Research Project Agency's Biological Technologies Office. Our team implemented a four-pronged independent verification and validation (IV&V) strategy including 1) a centralized data storage and exchange platform, 2) quality assurance and quality control (QA/QC) of data collection, 3) test and evaluation of performer models, and 4) an archival software and data repository. Results: Our IV&V plan was carried out with assistance from both the funding agency and participating teams of researchers. QA/QC of data acquisition aided in process improvement and the flagging of experimental errors. Holdout validation set tests provided an independent gauge of model performance. Conclusions: In circumstances that do not support a fully open approach to scientific criticism, standing up independent teams to cross-check and validate the results generated by primary investigators can be an important tool to promote reproducibility of results.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2025

Hourly Electricity Demand Profiles for Each County in the Contiguous United States

This dataset provides estimated hourly electricity demand for each county in the contiguous United States from 2016-2023. The demand profiles represent the sum of two components: (1) Weighted averages of reported hourly demand profiles for North American Electric Reliability Corporation balancing authority (BA) regions and subregions, scaled to match annual estimates of county-level retail sales and direct use of electricity and weighted by the estimated percentage of county load served by each BA region or subregion. (2) Weighted averages of modeled hourly, county- and sector-level distributed photovoltaic (DPV) capacity factor profiles, scaled to match annual estimates of on-site consumption of DPV-generated electricity for each county and weighted by the percentage of consumption attributable to each sector Annual county-level retail sales are estimated by aggregating utility-reported sales to the state level and allocating the results to counties according to each county's share of state population. Annual county-level direct use is calculated by aggregating power plant-reported direct use values. Annual county-level on-site consumption of DPV-generated electricity is estimated by aggregating utility-reported net metering data to determine the amount of DPV-generated electricity sold back to the grid for each state, subtracting those values from modeled state-level DPV generation estimates, and allocating the results to counties according to each county's share of statewide modeled DPV generation. The open-source Python code used to develop this dataset is available at "Historical Load Data Repository" link below.

14 SOLAR ENERGY↗
DOE OSTI2016

AutoSIM

AutoSIM enables several instances of an executable to run in parallel on a directory of input files and consolidate results, at scale on the world's fastest supercomputers. The software accomplishes this by balancing compression, transmission, and computation of input files, executables, and output files to allow scalable parallel runs of any executable. The software breaks-up simulations into a configurable number of "churns" or sequential instances of executable invocations on a configurable number of cores/node. It sends the executable, supporting files, and appropriate number of input files to each node for independent execution, and then transfers the executable results to a shared disk for post-processing analysis. This MPI-based code has been demonstrated using EnergyPlus, DOE's flagship whole-building simulation engine, but was architected to allow scalable runtime of any software for any application (e.g. sensitivity analysis and uncertainty quantification). The code has also been demonstrated to scale on the Titan supercomputer using 131,072 processors to run 524,288 EnergyPlus annual simulations, and write 45TB of data to disk in 68 mins 08 secs.

New, Joshua R↗
DOE OSTI2015

Comparative analysis of secretomes from Ectomycorrhizal fungi with an emphasis on small-secreted proteins

Fungi are major players in the carbon cycle in forest ecosystems due to the wide range of interactions they have with plants either through soil degradation processes by litter decayers or biotrophic interactions with pathogenic and ectomycorrhizal symbionts. Secretion of fungal proteins mediates these interactions by allowing the fungus to interact with its environment and/or host. Ectomycorrhizal (ECM) symbiosis independently appeared several times throughout evolution and involves approximately 80% of trees. Despite extensive physiological studies on ECM symbionts, little is known about the composition and specificities of their secretomes. In this study, we used a bioinformatics pipeline to predict and analyze the secretomes of 49 fungal species, including 11 ECM fungi, wood and soil decayers and pathogenic fungi to tackle the following questions: (1) Are there differences between the secretomes of saprophytic and ECM fungi? (2) Are small-secreted proteins (SSPs) more abundant in biotrophic fungi than in saprophytic fungi? and (3) Are there SSPs shared between ECM, saprotrophic and pathogenic fungi? We showed that the number of predicted secreted proteins is similar in the surveyed species, independently of their lifestyle. The secretome from ECM fungi is characterized by a restricted number of secreted CAZymes, but their repertoires of secreted proteases and lipases are similar to those of saprotrophic fungi. Focusing on SSPs, we showed that the secretome of ECM fungi is enriched in SSPs compared with other species. Most of the SSPs are coded by orphan genes with no known PFAM domain or similarities to known sequences in databases. Finally, based on the clustering analysis, we identified shared- and lifestyle-specific SSPs between saprotrophic and ECM fungi. The presence of SSPs is not limited to fungi interacting with living plants as the genome of saprotrophic fungi also code for numerous SSPs. As a result, ECM fungi shared lifestyle-specific SSPs likely involved in symbiosis that are good candidates for further functional analyses.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2019

ASC news highlight for HQ, 12/11/18

LLNL’s package management tool, Spack, has reduced software build times from weeks to hours for programmatic codes, and the tool’s popularity has grown beyond the NNSA laboratories. New workflow features and ongoing community-building efforts have led ECP and many top supercomputing sites worldwide to adopt Spack as their official deployment tool. With a community of over 330 Spack contributors, the maintenance burden for over 3,000 open source packages is now shared with other centers. This reduction of general ASC software maintenance costs frees up code development resources to address unique requirements specific to the stockpile stewardship mission.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Improving scalability of electronic structure code for molecular simulations in the presence of environment

A scalable density functional electronic code with Gaussian basis set, called UTEP-NRLMOL, is developed to perform simulations of molecular systems in the presence of the environment with particular attention to the memory requirements. In the electronic structure calculations, the memory and computation time are proportional to the number of atoms. Memory requirements for density functional calculations scale as N*N, where N is the number of atoms. While the recent advances in HPC offer platforms with large numbers of cores, the limited amount of memory available on a given node and poor scalability of the electronic structure codes hinder their efficient usage of these platforms. We have introduced new scaling and parallelization paradigms using MPI-3 shared-memory functionality combined with usage of sparse algebra and storage of matrices in sparse format. This extends the range of applicability of the UTEP-NRLMOL code to large systems over 10,000 atoms, or using up to 67,000 basis functions, and making use of HPC architectures using over 6,000 processors utilizing all available cores. We have also interfaced code with effective fragment potential and polarizable continuum model libraries. The code was used in simulations of several applications which are published in reputed scientific journals.

74 ATOMIC AND MOLECULAR PHYSICS↗
DOE OSTI2024

Parallel Simulation of Quantum Networks with Distributed Quantum State Management

Quantum network simulators offer the opportunity to cost-efficiently investigate potential avenues for building networks that scale with the number of users, communication distance, and application demands by simulating alternative hardware designs and control protocols. Several quantum network simulators have been recently developed with these goals in mind. As the size of the simulated networks increases, however, sequential execution becomes time-consuming. Parallel execution presents a suitable method for scalable simulations of large-scale quantum networks, but the unique attributes of quantum information create unexpected challenges. In this work, we identify requirements for parallel simulation of quantum networks and develop the first parallel discrete-event quantum network simulator by modifying the existing serial simulator SeQUeNCe. Our contributions include the design and development of a quantum state manager (QSM) that maintains shared quantum information distributed across multiple processes. We also optimize our parallel code by minimizing the overhead of the QSM and decreasing the amount of synchronization needed among processes. Using these techniques, we observe a speedup of 2 to 25 times when simulating a 1,024-node linear network topology using 2 to 128 processes. We also observe an efficiency greater than 0.5 for up to 32 processes in a linear network topology of the same size and with the same workload. We repeat this evaluation with a randomized workload on a caveman network. We also introduce several methods for partitioning networks by mapping them to different parallel simulation processes. We have released the parallel SeQUeNCe simulator as an open source tool alongside the existing sequential version.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Flexible Pilot Jobs Framework for Distributed High Throughput Computing

Experimental particle physics has been at the forefront of analyzing the world’s largest datasets for decades. The high-energy physics (HEP) community was among the first to develop suitable software and computing tools for this purpose. GlideinWMS is a Glidein-based workload management system whose purpose is to provide experiments like CMS at CERN, DUNE at Fermilab, and others, a way to access and efficiently use vast amounts of computing resources. This system wants to provide a simple way to submit jobs to a set of computing resources, that will be provided to users behind the scenes. Glideins are the pilot jobs executed on the worker nodes at the grid sites, performing operations such as hardware detection, environment setup, and error handling. After all these operations, they will launch the actual user job. Many grid sites are supported, such as shared clusters, Google CE, and AWS. My internship aimed to design and code a flexible pilot jobs framework that will replace the one used by GlideinWMS, developing a modular and flexible skeleton of the Glidein and adding further functionalities. My project also focused on the application of machine learning techniques as support to this management system.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2016

Early Experiences Writing Performance Portable OpenMP 4 Codes

In this paper, we evaluate the recently available directives in OpenMP 4 to parallelize a computational kernel using both the traditional shared memory approach and the newer accelerator targeting capabilities. In addition, we explore various transformations that attempt to increase application performance portability, and examine the expressiveness and performance implications of using these approaches. For example, we want to understand if the target map directives in OpenMP 4 improve data locality when mapped to a shared memory system, as opposed to the traditional first touch policy approach in traditional OpenMP. To that end, we use recent Cray and Intel compilers to measure the performance variations of a simple application kernel when executed on the OLCF s Titan supercomputer with NVIDIA GPUs and the Beacon system with Intel Xeon Phi accelerators attached. To better understand these trade-offs, we compare our results from traditional OpenMP shared memory implementations to the newer accelerator programming model when it is used to target both the CPU and an attached heterogeneous device. We believe the results and lessons learned as presented in this paper will be useful to the larger user community by providing guidelines that can assist programmers in the development of performance portable code.

Joubert, Wayne↗
DOE OSTI2020

OpenACC unified programming environment for GPU and FPGA multi-hybrid acceleration

Attached accelerators have been frequently used in recent High Per- formance Computing (HPC) systems because of their high performance/power ratio. In particular, the Graphics Processing Unit (GPU) is the most popu- lar accelerator owing to its high peak FLOPS performance and high memory bandwidth supported by HBM2, etc. However, the performance of GPU depends highly on a large degree of SIMD parallelism and has difficulty sustaining a high performance on programs with frequent branch operations or a partially low degree of parallelism.By contrast, a Field Programmable Gate Array (FPGA) has received attention as a different type of accelerator than GPU as a fully reconfigurable processor fitting the target applications. The high performance of FPGA is mainly provided by a pipelined operation and optimized circuit suitable for any operation even with frequent conditional branches. We have been focusing on the flexibility of FPGA to compensate for the weakness of GPU. We believe that the coupling of GPU with FPGA can result in one of the most powerful accelerating platforms available.However, the program coding of GPU and FPGA coupling can be quite difficult for application users. Traditionally, CUDA by NVIDIA has been the most popular programming language with the largest share of GPUs used in HPC, whereas a hardware description language such as Verilog HDL has been used in FPGA programming. OpenCL coding has recently become available even on high-end FPGAs. Moreover, several recent studies have also enabled the OpenACC coding for use in FPGA. In this study, we provide a unified programming system based on OpenACC for a platform equipped with both GPU and FPGA aiming at the next-generation accelerated supercomputer framework. Our programming environment is called Multi-Hybrid OpenACC Translator (MHOAT), and in this paper, we describe the basic concept and prototype system of MHOAT based on an evaluation on the amount of coding required and the performance of a hybrid multi-device accelerated system.

Tsunashima, Ryuta↗
DOE OSTI2024

An Energy Codes Gap Analysis Field Study in the Southwest (Final Technical Report)

The primary goal of the “An Energy Codes Gap Analysis Field Study in the Southwest” project was to improve housing through strengthened building energy code implementation in two states, Colorado and Nevada, leading to greater energy and utility bill savings for households. Colorado is a home rule state. Nevada adopts a statewide code, but cities and counties must also adopt the code, making the state function like a home rule state. To broaden the impact of the project, an additional goal was to share and amplify project experiences, findings, and lessons that can be applied in other states. To achieve these goals, the project’s key objectives were to examine how residential energy codes are implemented in partner states; to use these findings to strengthen codes implementation through enhanced education, training, and outreach; and to enhance energy code technical assistance capabilities in the two partner states. In both partner states, the project led to the creation of a quantitative baseline understanding of compliance with high-impact energy code requirements, such as foundation, wall, and ceiling R-value; window U-factor and SHGC; envelope air tightness; duct tightness; and high-efficiency lighting. It then applied this baseline information to directly inform training in geographic areas with high levels of building and construction activity. The project has placed Colorado and Nevada among 23 states across the country that have conducted a single-family residential field study based on the established U.S. Department of Energy (DOE) methodology since 2014. It has also helped each State Energy Office test models for partnership- and relationship-building among government agencies and key stakeholder groups, including the State Energy Office and relevant code administration and professional licensing agencies, building officials, designers, builders, trades and utilities. At the national level, the project helped educate and inform State and Territory Energy Offices on a replicable methodology to assess energy code construction practices, effective stakeholder engagement strategies, and tailored energy code education and training. These findings are particularly pertinent for home-rule states, such as Colorado and Nevada, which rely on local government awareness and action to advance and implement building energy codes. By raising awareness of code adoption and compliance, this project has helped to equip each partner state with data on code compliance, strategies, education, and partnership models to advance building energy codes. This in turn will have an important impact on new construction practices in Colorado and Nevada, leading to more energy efficient and affordable housing. Currently, more than 62 percent of Colorado’s population lives in one of the 55 jurisdictions that have adopted the 2021 IECC to date, while nearly 95 percent of the state's population lives in one of the 208 jurisdictions that have adopted a code at 2015 IECC levels or higher. In Nevada, the 2018 IECC was adopted by the Governor’s Office of Energy in July 2018. By July 2020, 47 percent of the adopted energy code in the state was 2018 IECC which covers 96.5 percent of Nevada’s population. The training materials funded by this project focus on the 2021 IECC and will continue to be relevant in the two states as additional jurisdictions update their building energy codes.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2026

User’s Manual for RESRAD-RDD&IND Code Version 2: Vol. 2—User’s Guide for RESRAD-RDD&IND Code

Version 2.0 of the RESRAD-RDD&IND computer code is designed to support the implementation of protective action guides (PAGs) after a nuclear emergency incident including a radiological dispersal device (RDD) and/or an improvised nuclear device (IND) incident (EPA 2017). Eight different group types, addressing various decisions, are available for selection. The RESRAD-RDD&IND code calculates radiological doses, stay times, etc., for the selected group that the user wishes to focus on. (That is, the results for all the groups are not calculated simultaneously, and the input for those other groups do not matter, although some parameter values are shared between groups.) Version 2.0 has a user-friendly interface so that the RESRAD-RDD&IND code can be used with minimal training. For example, the user can select the major characteristics of the problem-event type, source term, and decision type from the left side of the interface and then calculate the results with the default assumptions for the exposure scenarios. More in-depth analysis would include specifying site-specific exposure scenario characteristics in the right side of the interface. The procedures for data entry and results viewing are self-explanatory. This is because common window maneuvering features and text instructions were incorporated in the interface design. General and context-specific help are available to aid users entering parameter values, as well. The RESRAD-RDD&IND computer code gives the user the option to select either an RDD or IND incident for analysis. For an RDD event analysis, 11 radionuclides (Am-241, Cf-252, Cm-244, Co-60, Cs-137, Ir-192, Po-210, Pu-238, Pu-239, Ra-226, and Sr-90) are included. These 11 radionuclides are the radionuclides most likely used for an RDD. More than 90 radionuclides can be selected for an IND event analysis. Initial default concentrations are provided for 44 radionuclides for a uranium-fueled IND event. These 44 radionuclides are those that would contribute significantly to the radiation dose associated with a uranium-fueled bomb detonation. The radionuclides generated from ingrowth of these 44 initial radionuclides are also automatically included in the analysis. Pu-239, Cs-134m, Ru-105, and Rb-89 and their progeny can be selected for analysis if they are detected and their concentrations are determined. This user’s guide, which is Volume 2 of the User’s Manual for RESRAD-RDD&IND Code Version 2, provides instructions to users on how to install the RESRAD-RDD&IND code, navigate the interface, and use the various features, including those discussed above, to set up an analysis and view/print the results in text outputs. Volume 1 of the User’s Manual for RESRAD-RDD&IND Code Version 2 (Yu et al. 2026), which contains descriptions of the methodology and theoretical basis for dose modeling and the mathematical equations implemented in the code, can be accessed and viewed through the Help menu in the code or can be downloaded from the RESRAD website (https://resrad.evs.anl.gov).

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2019

Performance optimization, modeling and analysis of sparse matrix-matrix products on multi-core and many-core processors

Sparse matrix-matrix multiplication (SpGEMM) is a computational primitive that is widely used in areas ranging from traditional numerical applications to recent big data analysis and machine learning. Although many SpGEMM algorithms have been proposed, hardware specific optimizations for multi- and many-core processors are lacking and a detailed analysis of their performance under various use cases and matrices is not available. In this work, we firstly identify and mitigate multiple bottlenecks with memory management and thread scheduling on Intel Xeon Phi (Knights Landing or KNL). Specifically targeting many-core processors, we develop a hash-table-based algorithm and optimize a heap-based shared-memory SpGEMM algorithm. We examine their performance together with other publicly available codes. Different from the literature, our evaluation also includes use cases that are representative of real graph algorithms, such as multi-source breadth-first search or triangle counting. Our hash-table and heap-based algorithms are showing significant speedups from libraries in the majority of the cases while different algorithms dominate the other scenarios with different matrix size, sparsity, compression factor and operation type. We wrap up in-depth evaluation results and make a recipe to give the best SpGEMM algorithm for target scenario. We build the performance model for hash-table and heap-based algorithms, which supports the recipe. A critical finding is that hash-table-based SpGEMM gets a significant performance boost if the nonzeros are not required to be sorted within each row of the output matrix. Finally, we integrate our implementations into a large-scale protein clustering code named HipMCL, accelerating its SpGEMM kernel by up to 10X and achieving an overall performance boost for the whole HipMCL application by 2.6X.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

Soluble limit and criticality of fermions in $\mathbb Z_2$ gauge theories

Quantum information theory and strongly correlated electron systems share a common theme of macroscopic quantum entanglement. In both topological error correction codes and theories of quantum materials (spin liquid, heavy fermion and high-$T_c$ systems), entanglement is implemented by means of an emergent gauge symmetry. Inspired by these connections, in this paper we introduce a simple model for fermions moving in the deconfined phase of a $\mathbb Z_2$ gauge theory by coupling Kitaev's toric code to mobile fermions. This permits us to exactly solve the ground state of this system and map out its phase diagram. Reversing the sign of the plaquette term in the toric code permits us to tune the ground state between an orthogonal metal and an orthogonal semimetal in which gapless quasiparticles survive despite a gap in the spectrum of original fermions. The small-to-large Fermi surface transition between these two states occurs in a stepwise fashion with multiple intermediate phases. By using a diagrammatic technique, we are able to explore physics beyond the integrable point to examine various instabilities of the deconfined phase and to derive the critical theory at the transition between deconfined and confined phases. We outline how the fermionic toric code can be implemented as a quantum circuit, thus providing an important link between quantum materials and quantum information theory.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
DOE OSTI2020

The genome of jojoba ( Simmondsia chinensis ): A taxonomically isolated species that directs wax ester accumulation in its seeds

Seeds of the desert shrub, jojoba ( Simmondsia chinensis ), are an abundant, renewable source of liquid wax esters, which are valued additives in cosmetic products and industrial lubricants. Jojoba is relegated to its own taxonomic family, and there is little genetic information available to elucidate its phylogeny. Here, we report the high-quality, 887-Mb genome of jojoba assembled into 26 chromosomes with 23,490 protein-coding genes. The jojoba genome has only the whole-genome triplication (γ) shared among eudicots and no recent duplications. These genomic resources coupled with extensive transcriptome, proteome, and lipidome data helped to define heterogeneous pathways and machinery for lipid synthesis and storage, provided missing evolutionary history information for this taxonomically segregated dioecious plant species, and will support efforts to improve the agronomic properties of jojoba.

Sturtevant, Drew↗
DOE OSTI2018

High-Performance Sparse Matrix-Matrix Products on Intel KNL and Multicore Architectures

Sparse matrix-matrix multiplication (SpGEMM) is a computational primitive that is vastly used in areas ranging from traditional numerical applications to recent big data analysis and machine learning. While many SpGEMM algorithms have been proposed, hardware specific optimizations for multi- and many-core processors are lacking and a detailed analysis of their performance under various use cases and matrices is not available. We firstly identify and mitigate multiple bottlenecks with memory management and thread scheduling on Intel Xeon Phi (Knights Landing or KNL). Specifically targeting multi- and many-core processors, we develop a hash-table-based algorithm and optimize a heap-based shared-memory SpGEMM algorithm. We investigate their performance together with other publicly available codes. Different from the literature, our evaluation also includes use cases that are representative of real graph algorithms, such as multi-source breadth-first search or triangle counting. Our hash-table and heap-based algorithms are showing significant speedups from libraries in the majority of the cases while different algorithms dominate the other scenarios with different matrix size, sparsity, compression factor and operation type. We wrap up in-depth evaluation results and make a recipe to give the best SpGEMM algorithm for target scenario. In conclusion, a critical finding is that hash-table-based SpGEMM gets a significant performance boost if the nonzeros are not required to be sorted within each row of the output matrix.

97 MATHEMATICS AND COMPUTING↗