Papers and Thesis:
Ledur, C. , Griebler, D. , Manssour, I. , Fernandes, L. G. . Towards a Domain-Specific Language for Geospatial Data Visualization Maps with Big Data Sets.In: ACS/IEEE International Conference on Computer Systems and Applications (AICCSA'15). 2015. [DOI] [PDF]
Ledur, C. L. ; Griebler, D. J. ; Manssour, I. H. ; Fernandes, L. G. . Uma Linguagem Específica de Domínio com Geração de Código Paralelo para Visualização de Grandes Volumes de Dados. In: 15a ERAD - Escola Regional de Alto Desempenho, 2015, Gramado. Anais da 15a ERAD, 2015. p. 139-140. [PDF]
GMaVis Research Context:
GMaVis was built as a result of a master thesis. GMaVis and this work contributes to a framework built at GMAP research group. It is the result of a set of researches and works developed in the research group that has as main objective to facilitate the creation and implementation of parallel applications. GMaVis was built on top of this framework, at the application level, generating SPar annotations and using its compiler to generate a parallel data preprocessor.
SPar is a domain-specific language that allows the parallelism of code using annotations and the implementation of parallel processing based on stream parallelism. SPar was proposed to address stream processing applications. It is a C++ embedded domain-specific language (DSL) for expressing stream parallelism by using standard C++11 attribute annotations. It introduces high-level parallel abstractions for developing stream based parallel
programs as well as reducing sequential source code rewriting. Spar allows C++ developers to express stream parallelism by using the standard syntax grammar of the host language. It also enables minimal sequential code rewriting thus reducing the effort needed to program the parallel application. Additionally, SPar provides flexibility for annotating the C++ sequential code in different ways. In the DSL Generation Engine, Cincle, a compiler infrastructure for new C/C++ language extensions enables SPar compiler to generate code to FastFlow and MPI (Message Passing Interface) that takes advantage of different architecture systems. Therefore, SPar simplified the parallelization of the data preprocessor module by enabling GMaVis to compile the same code for both parallel or sequential execution by just modifying a compiler argument. GMaVis used the same code as the sequential version annotated in order to generate parallel version with SPar compiler. Also, it enabled GMaVis to easily abstract the parallel programming completely from users. Thus, domain users will not have to worry about creating parallel programming code to speed up data processing during visualization creation.