2018 IEEE 25th International Conference on High Performance Computing Workshops (HiPCW)最新文献

{"title":"DInEMMo: Decentralized Incentivization for Enterprise Marketplace Models","authors":"Ashwini Marathe, K. Narayanan, Avantika Gupta, P. Manoj","doi":"10.1109/HIPCW.2018.8634320","DOIUrl":"https://doi.org/10.1109/HIPCW.2018.8634320","url":null,"abstract":"Today, Machine learning (ML) / Artificial Intelligence (AI) has revolutionized the way through which data is perceived. Enterprises are using ML models to gain insights from the data and build applications of highest quality and accuracy. In this process, they are trying to seek more data to derive robust conclusions. However, relevant data are privately held and resides with an organization's premise which thwarts the development of accurate models. Decentralized AI has become an attractive technological trend for enterprises as it ensures model improvement and creates a demand for them through a marketplace. Nonetheless, its potential can be unleashed if there is a massive user participation enabled through fair rewards to its contributors. Motivated by these observations, in this paper, we present DInEMMo, a solution that is built on the convergence of decentralized AI and Blockchain. DInEMMo is enabled with configurable smart contracts with the following features: (1) represent the ML model and use case attributes, (2) generation of models (new / enhanced) based on user input, (3) compute the price of the ML model based on the user policy, and (4) calculate the incentives to the model's owner and co-contributors. Using these features, we qualitatively evaluate the relevancy of the system for the use case on Medical Diagnostics and show the significance of domain specific properties in rewarding the contributors and further, determining the model price.","PeriodicalId":401060,"journal":{"name":"2018 IEEE 25th International Conference on High Performance Computing Workshops (HiPCW)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124878299","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Smart Contracts for Multiagent Plan Execution in Untrusted Cyber-Physical Systems","authors":"Anshu Shukla, S. Mohalik, R. Badrinath","doi":"10.1109/HIPCW.2018.8634034","DOIUrl":"https://doi.org/10.1109/HIPCW.2018.8634034","url":null,"abstract":"Intelligent Cyber-physical systems can be modeled as multi-agent systems with planning capability to impart adaptivity for changing contexts. In such multi-agent systems, the protocol for plan execution must result in the proper completion and ordering of actions in spite of their distributed execution. However, in untrusted scenarios, there is a possibility of agents not respecting the protocol either due to faults or due to malicious reasons thereby resulting in plan failure. In order to prevent such situations, we propose to implement the execution of agents through smart contracts. This points to a generic architecture seamlessly integrating intelligent planning-based CPS and smart-contracts.","PeriodicalId":401060,"journal":{"name":"2018 IEEE 25th International Conference on High Performance Computing Workshops (HiPCW)","volume":"57 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127076209","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Acceleration of a 3D Immersed Boundary Solver Using OpenACC","authors":"Apurva Raj, Somnath C. Roy, N. Vydyanathan, Bharatkumar Sharma","doi":"10.1109/HIPCW.2018.8634138","DOIUrl":"https://doi.org/10.1109/HIPCW.2018.8634138","url":null,"abstract":"Immersed-boundary methods (IBM) have been constantly gaining popularity and are increasingly expanding to new areas of applications in computational mechanics since last three decades due to the potentials of their application in modeling complex multiphysics phenomena which involves flow over complex and moving boundaries. The specific advantages of an immersed boundary method are due to its accuracy and simplicity. As this method uses a fixed structured Cartesian mesh, the complex grid generation processes can be fully avoided whereas the complex/moving boundary is described using another surface mesh. The computational overheads in an immersed boundary implementation can be very high due to expensive search and interpolation steps through which the effects of the boundary conditions on the surface mesh are translated to the fixed Cartesian volume mesh. Therefore, computationally efficient numerical implementation of an IBM solver is of extreme importance to researchers. This paper presents an accelerated discrete finite difference based immersed boundary (IB) solver that is used to study the external flow behavior around complex geometries. The flow is assumed to be incompressible. The immersed boundary solver is parallelized using OpenACC for quick acceleration with minimal code changes and to ensure performance portability across both GPUs and multicore CPUs. Our experimental results indicate that the OpenACC-based IB solver run on a NVIDIA Tesla P100 GPU is 21x faster than the sequential legacy solver and is 3.3x faster than the OpenACC-based IB solver run on a dual socket Intel Xeon Gold 6148, 20 core CPU. The recirculation lengths obtained for Reynolds numbers of 20 and 40 and the Strouhal number for Reynolds number 100, for a standard flow visualization problem over a fixed cylinder, are in accordance with the reported data in available literature, thereby validating the accuracy of the parallel solver. We also analyze the performance of the accelerated solver on different GPU architectures: Kepler, Pascal and Volta.","PeriodicalId":401060,"journal":{"name":"2018 IEEE 25th International Conference on High Performance Computing Workshops (HiPCW)","volume":"194 5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116785519","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Optimizing the Fast Fourier Transform Using Mixed Precision on Tensor Core Hardware","authors":"Anumeena Sorna, Xiao-he Cheng, E. D'Azevedo, Kwai Wong, S. Tomov","doi":"10.1109/HIPCW.2018.8634417","DOIUrl":"https://doi.org/10.1109/HIPCW.2018.8634417","url":null,"abstract":"The Fast Fourier Transform is a fundamental tool in scientific and technical computation. The highly parallelizable nature of the algorithm makes it a suitable candidate for GPU acceleration. This paper focuses on exploiting the speedup due to using the half precision multiplication capability of the latest GPUs' tensor core hardware without significantly degrading the precision of the Fourier Transform result. We develop an algorithm that dynamically splits the input single precision dataset into two half precision sets at the lowest level, uses half precision multiplication, and recombines the result at a later step. This work paves the way for using tensor cores for high precision inputs.","PeriodicalId":401060,"journal":{"name":"2018 IEEE 25th International Conference on High Performance Computing Workshops (HiPCW)","volume":"67 1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125849903","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Perforated Bluff-Body Wake Simulations: Influence of Aspect Ratio","authors":"Abhinav Singh, V. Narasimhamurthy","doi":"10.1109/HIPCW.2018.8634018","DOIUrl":"https://doi.org/10.1109/HIPCW.2018.8634018","url":null,"abstract":"Parallel computations of flow past a perforated plate of porosity 25% at Reynolds number 250 (based on plate width, d and inflow velocity, Uo) is carried out. The effect of aspect ratio is studied with different span-wise lengths of the domain (1d, 3d and 6d). Present results revealed that an aspect ratio of 6d is required to capture the transient wake dynamics. It was found that statistical quantities stemming from aspect ratio 3d and 6d cases agree with each other, though the dynamical behavior of the wake is very different. The signature period doubling effects associated with short constrained domains were visible in the 1d and 3d aspect ratio cases. Enforcing periodic boundary condition along the short span-wise domains may thus adversely affect the flow.","PeriodicalId":401060,"journal":{"name":"2018 IEEE 25th International Conference on High Performance Computing Workshops (HiPCW)","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122851169","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Data Science Techniques to Improve Accuracy of Provider Network Directory","authors":"Priya Kandasamy, Divya Raji, Arun Sundararaman","doi":"10.1109/HIPCW.2018.8634423","DOIUrl":"https://doi.org/10.1109/HIPCW.2018.8634423","url":null,"abstract":"Trivial or tactical as it may appear, yet, Provider data inaccuracy continues to be a major challenge in healthcare industry. With about 250 key attributes per provider and roughly 500K providers in USA, this translates to maintaining current and correct values for a whopping 12.5 M attributes dataset that is very dynamic and volatile. Inaccuracy in this dataset implies 2 major adverse consequences; a) Regulatory penalties ranging from few thousand dollars to few million dollars and b) potential member attrition due to member dissatisfaction, triggered by increased waiting time, delay in accessing the medical service, efforts wasted on reaching out to incorrect provider etc. Many of the current solutions carry limitations such as lack of centralized storage, data latency issues and non-standardized questionnaire to capture provider update etc. This paper introduces an innovative approach that addresses these limitations using Predictive Analytics and Intake Scoring techniques. Rooted in Data Science, the proposed ensemble model combines the advantages of individual prediction models such as Logistic Regression, Random Forest, Neural Network and XgBoost. This automated approach also brings down the dependency on external systems and automatically updates the database, keeping it up to date. A detailed analysis of results from work carried out using this innovative approach are discussed at length and the paper concludes with directions for future work.)","PeriodicalId":401060,"journal":{"name":"2018 IEEE 25th International Conference on High Performance Computing Workshops (HiPCW)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123720589","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Three Dimensional Pseudo-Spectral Compressible Magnetohydrodynamic GPU Code for Astrophysical Plasma Simulation","authors":"R. Mukherjee, R. Ganesh, V. Saini, U. Maurya, N. Vydyanathan, Bharatkumar Sharma","doi":"10.1109/HIPCW.2018.8634104","DOIUrl":"https://doi.org/10.1109/HIPCW.2018.8634104","url":null,"abstract":"This paper presents the benchmarking and scaling studies of a GPU accelerated three dimensional compressible magnetohydrodynamic code. The code is developed keeping an eye to explain the large and intermediate scale magnetic field generation is cosmos as well as in nuclear fusion reactors in the light of the theory given by Eugene Newman Parker. The spatial derivatives of the code are pseudo-spectral method based and the time solvers are explicit. GPU acceleration is achieved with minimal code changes through OpenACC parallelization and use of NVIDIA CUDA Fast Fourier Transform library (cuFFT). NVIDIA's unified memory is leveraged to enable oversubscription of the GPU device memory for seamless out-of-core processing of large grids. Our experimental results indicate that the GPU accelerated code is able to achieve upto two orders of magnitude speedup over a corresponding OpenMP parallel, FFTW library based code, on a NVIDIA Tesla P100 GPU. For large grids that require out-of-core processing on the GPU, we see a 7x speedup over the OpenMP, FFTW based code, on the Tesla P100 GPU. We also present performance analysis of the GPU accelerated code on different GPU architectures - Kepler, Pascal and Volta.","PeriodicalId":401060,"journal":{"name":"2018 IEEE 25th International Conference on High Performance Computing Workshops (HiPCW)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115279499","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}