english Icono del idioma   español Icono del idioma  

Please use this identifier to cite or link to this item: https://hdl.handle.net/20.500.12008/36663 How to cite
Title: Refactoring an electric-market simulation software for massively parallel computations
Authors: Seveso, Franco
Marichal, Raúl
Dufrechou, Ernesto
Ezzatti, Pablo
Type: Preprint
Keywords: Coarse-grained parallelism, Electric energy generation, Stochastic dynamic programming, Memory usage
Issue Date: 2022
Abstract: In the last two decades, Uruguay has been immersed in the process of significantly changing its energy generation matrix, especially by the introduction of wind and solar sources. In this context, SimSEE, a simulation and optimization software designed to help decision-making in generating and distributing electrical energy, is extensively used. The design of this tool is conceived for conventional CPUs and follows a sequential execution paradigm. This paper focuses on a refactoring of SimSEE that enables leveraging massively-parallel hardware platforms, seeking to adapt the tool for the increasing size and complexity of Uruguay’s electric market. We extend our previous ideas about reorganizing the software architecture to exploit the parallelism in each time-step of Sim-SEE’s simulation. In more detail, we present two variants following this parallelism pattern, a straightforward parallel version that requires replicating the used memory and a variant that implies limited performance restrictions but requires a minimal memory overhead.
Description: Latin America High Performance Computing Conference, CARLA 2022, Porto Alegre, Brazil.
Sponsors: Agencia Nacional de Investigación e Innovación. Proyecto ANII FSE_1_2018_1_153060 Aceleración del SimSEE utilizando GPUs (SimSEE-MP).
Citation: Seveso, F., Marichal, R., Dufrechou, E. y otros. Refactoring an electric-market simulation software for massively parallel computations [Preprint] Publicado en: Latin America High Performance Computing Conference, CARLA 2022, Porto Alegre, Brazil, Sep. 26-30, 2022, pp.190-204.
License: Licencia Creative Commons Atribución - No Comercial - Sin Derivadas (CC - By-NC-ND 4.0)
Appears in Collections:Reportes Técnicos - Instituto de Computación

Files in This Item:
File Description SizeFormat  
SMDE22.pdfPreprint536,09 kBAdobe PDFView/Open


This item is licensed under a Creative Commons License Creative Commons