Visiting professor engaged in data processing activities
Gabriele Mencagli, from Università di Pisa, develops research at School of Technology
Photo: Mariana Haupenthal
The professor of the Università di Pisa, Italy, Gabriele Mencagli, came to PUCRS as a visiting professor under the Institutional Project of Internationalization (PUCRS-PrInt) earlier in September. The professor is works at the Department of Computer Science of the Italian institution and is working on a number of initiatives at the School of Technology.
Mencagli agenda includes lectures, graduate courses and other activities with the research group Modeling and Parallel Applications, coordinated by Dr Luiz Gustavo Leão Fernandes.
A number of initiatives are expected to be produced from this visit: publication of co-authored articles in international journals, as well as a research internship of a gradaute student in the Graduate Program in Computer Science (PPGCC) Adriano Jose Vogel, in Dec 2019. Adriano will spend twelve months at the Università de Pisa, where he will be welcomed by the research group Parallel Programming Models Group, of which Dr Gabriele is a member.
On Sept 9, Mencagli delivered the lecture WindFlow: a C++17 parallel library for data stream processing on multicores. The conference introduced the audience to WindFlow, a parallel library for efficient flow processing in modern multi-core systems. The library uses the parallel building blocks provided by FastFlow, a parallel programming framework developed at the University of Pisa. The basic FastFlow components and their semantics, the C ++ 17-based WindFlow Application Programming Interface (API) and its parallel execution system were presented.
Data Flow Processing
Photo: Mariana Haupenthal
Mencagli is also teaching the course on Data Stream Processing (from the parallelism perspective), to the students in the Graduate Program in Computer Science. In a scenario where everyone is faced with a growing number of devices capable of producing or generating large amounts of data all over the world, the course aims to provide a solid background to the research into efficient processing techniques. Programming comprises an introduction to the theory behind streaming data processing, an overview of the best known streaming processing systems (e.g. Apache Storm, Apache Flink and Spark Streaming) and the latest findings achieved by the scientific community in terms of optimizations, standards and cost models.
In addition, the visit is expected to result in the development of two new joint research papers, as follows: Towards novel stream processing frameworks and A complete benchmark suite in C++ for parallel NAS. Mencagli is working under the cooperation project Extracting knowledge from Big Data, under the priority area Technology and Biodiversity: Sustainability, Energy and the Environment.