The 25th edition of the Brazilian Symposium on Programming Languages (SBLP) took place in virtual format from September 27 to October 1, 2021, promoted by the Brazilian Computer Society (SBC). SBLP 2021 is a forum for students, researchers, and practitioners to present and discuss ideas and innovations in the design, definition, analysis, implementation, and practical use of programming languages. The symposium is part of the Brazilian Conference on Software: Practice and Theory (CBSoft). In addition to SBLP, CBSoft annually integrates three other traditional events promoted by the Brazilian community on Software Engineering: the 35th Brazilian Symposium on Software Engineering (SBES), the 15th Brazilian Symposium on Software Components, Architectures, and the Reuse (SBCARS), and 6th Brazilian Symposium on Systematic Automated Software Testing (SAST).
GMAP Master’s student Júnior Löff presented in SBLP 2021 the paper entitled “High-Level Stream and Data Parallelism in C++ for Multi-Cores”. The work was advised by Dr. Luiz Gustavo Fernandes, and co-advised by Dr. Dalvan Griebler. In addition, it had the collaboration of GMAP master’s student Renato B. Hoffmann.
The paper presents a high-level solution for expressing data and stream parallelism in C++ programs. Stream processing applications have become popular with the spread of sensors, IoT devices, user data etc. Applications that belong to this ecosystem usually require real-time results, which has become a challenge as the amount of data generated increases. In this work, a solution for expressing stream and data parallelism using a single high-level language has been introduced. To make this possible, a new compiler algorithm was implemented and new transformation rules were created to improve the intelligence of this compiler. Results showed that using our language and compiler, the programmer is able to parallelize a sequential program with few lines of code. In addition, the performance obtained is compared to manually parallelized versions. This research is promising because a big step has been taken in popularizing the writing of parallel code, which is still rarely used in the software industry. In the future, it is expected to be feasible to program, debug, and maintain code written in parallel, positively impacting the efficient use of computational resources, improving aspects such as energy efficiency and sustainability.
In 2021, the SBLP awards the two best papers, one for each category: student paper and non-student paper. The paper presented by Júnior Löff received the best student paper award at SBLP 2021.
Júnior tells us about his experience at the event:
“The presentation at the event was challenging, as it was my first presentation in English. However, everything went well and the work raised interesting discussions in the community. This work was one of several that have been developed at GMAP/PUCRS and that have contributed to the SPar language, a framework for parallel programming created in our research group. Is very satisfying to see the importance that the SPar results are achieving and the possible impact for the future of software engineering on writing parallel code. This award reinforces our research goal, to bring results that contribute to different areas of society.”
Júnior Löff
More information about SBLP 2021 can be found on the official website.
