GPU algorithms for Efficient Exascale Discretizations

2021-12-01
Abdelfattah, Ahmad
Barra, Valeria
Beams, Natalie
Bleile, Ryan
Brown, Jed
Camier, Jean-Sylvain
Carson, Robert
Chalmers, Noel
Dobrev, Veselin
Dudouit, Yohann
Fischer, Paul
Karakuş, Ali
Kerkemeier, Stefan
Kolev, Tzanio
Lan, Yu-Hsiang
Merzari, Elia
Min, Misun
Phillips, Malachi
Rathnayake, Thilina
Rieben, Robert
Stitt, Thomas
Tomboulides, Ananias
Tomov, Stanimire
Tomov, Vladimir
Vargas, Arturo
Warburton, Tim
Weiss, Kenneth
In this paper we describe the research and development activities in the Center for Efficient Exascale Discretization within the US Exascale Computing Project, targeting state-of-the-art high-order finite-element algorithms for high-order applications on GPU-accelerated platforms. We discuss the GPU developments in several components of the CEED software stack, including the libCEED, MAGMA, MFEM, libParanumal, and Nek projects. We report performance and capability improvements in several CEED-enabled applications on both NVIDIA and AMD GPU systems.
Parallel Computing

Suggestions

CLOUDGEN: Workload generation for the evaluation of cloud computing systems CLOUDGEN: Bulut Bilişim Sistemlerinin Başarim Deǧerlendirmesi icin Iş Yuku Uretimi
Koltuk, Furkan; Yazar, Alper; Schmidt, Şenan Ece (2019-04-01)
In this paper, we propose CLOUDGEN workflow that produces synthetic workloads for Infrastructure and Platform as a Service for the evaluation of resource management approaches in cloud computing systems. To this end, CLOUDGEN systematically processes and clusters records in a given workload trace and fits distributions for different workload parameters within the clusters. Different than the previous work, clustering is carried out to produce different virtual machine types for achieving models that are sui...
Electromagnetic Target Classification using time frequency analysis and neural networks
Sayan, Gönül; Leblebicioğlu, Mehmet Kemal (Wiley, 1999-04-01)
This paper demonstrates the feasibility and advantages of using a self-organizing map (SOM)-type neural network classifier for electromagnetic target recognition. The classifier is supported by a novel feature extraction unit in which the Wigner distribution (WD), a time-frequency representation, is utilized for the extraction of natural-resonance-related energy feature vectors from scattered fields. The proposed target classification technique is tested for a set of canonical targets, displaying an excelle...
ACCLOUD-MAN - Power efficient resource allocation for heterogeneous clouds ACCLOUD-MAN - Heterojen bulutlarda güç etkin kaynak atamasi
Ekici, Nazim Umut; Schmidt, Klaus Werner; Yazar, Alper; Schmidt, Şenan Ece (2019-04-01)
In this paper we propose ACCLOUD-MAN, a novel resource manager for heterogeneous cloud data centers. In heterogeneous clouds a user request can be satisfied with more than one physical resource alternative. Resource manager must decide which resource alternative will be chosen, along with the decision of the server the request will be assigned to. ACCLOUD-MAN's resource management objective is to reduce the power consumption of the cloud. Manager is modeled as an Integer Linear Problem and is implemented on...
High efficiency combined - cycle gas polygenerator for ecological local generation (HEGEL)
Pınarcıoğlu, Mehmet Melih(2009-4-30)
Objective is to develop, demonstrate and assess an innovative, high efficiency concept of micro-cogeneration system applied to a real demand site under real operating conditions. The application concept is based on a combined cycle architecture (Combi system) constituted by two integrated cogenerators powered by different prime movers: an innovative reciprocating engine cogenerator and a Rankine engine system (bottoming cycle) operated on the exhaust gases of the reciprocating engine. The location will b...
Neural network based beamforming for linear and cylindrical array applications
Güreken, Murat; Dural Ünver, Mevlüde Gülbin; Department of Electrical and Electronics Engineering (2009)
In this thesis, a Neural Network (NN) based beamforming algorithm is proposed for real time target tracking problem. The algorithm is performed for two applications, linear and cylindrical arrays. The linear array application is implemented with equispaced omnidirectional sources. The influence of the number of antenna elements and the angular seperation between the incoming signals on the performance of the beamformer in the linear array beamformer is studied, and it is observed that the algorithm improves...
Citation Formats
A. Abdelfattah et al., “GPU algorithms for Efficient Exascale Discretizations,” Parallel Computing, vol. 108, pp. 0–0, 2021, Accessed: 00, 2021. [Online]. Available: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85115934300&origin=inward.