RISC Project: let's build HPC together

"If you don't compute, you don't compete", that is how Mateo Valero, a seasoned HPC expert and the director of the Barcelona Supercomputing Center, summarizes the fundamental need to develop HPC infrastructure and develop science, engineering, government, and industry in the modern society. However, for some regions - Latin America among them - it is hard to keep up with the relentless development of powerful machines in other latitudes. To bridge this gap, sometimes it is necessary to build a network of stakeholders that share a common goal, agree to team up, and work in conjunction to provide all the actors in society with the right computing infrastructure that make all of them competitive.

That goal was among the main drivers behind the RISC project. A group of European and Latin American organizations worked together to understand the limitations, actors, and opportunities of HPC in Latin America. During more than 2 years, these groups worked in identifying the key components of a successful HPC initiative in Latin America. Among the specific goals of RISC project were: a) strengthen the cooperation between Europe and Latin America through the construction of a community to increase capacity, awareness, networking, and training; b) asses the drivers and needs of HPC in Latin America; c) make a roadmap for HPC development in Latin America; d) identify research clusters in Latin America; and e) paving the road for a dialogue between the research communities and the policy makers.

The RISC project identified research communities that were hungry for computing power, and that would benefit from access to HPC resources. The main areas pinned down by the project were: a) computational biology, b) oil exploration modeling, c) modeling and simulation of natural disasters, d) innovation in technology.

Another important contribution of the RISC project was a green paper on recommendations to develop HPC in Latin America. The paper suggests reducing the gap in HPC infrastructure in the region compared to the global leaders by consolidating a joint infrastructure in the whole region. Such infrastructure should follow a hierarchy with three layers of HPC resources (national, regional, and local), with the first layer of most powerful supercomputers being shared among users in the different countries. A more urgent recommendation in the paper is to obtain access to leadership facilities in other places to avoid the exodus of well trained and competitive scientists and engineers. In addition, the paper stresses the importance of developing training programs, constructing open-source software, pushing forward the discussion on policy making for HPC infrastructure, consolidating cooperation agreements with other entities, and more.

To know more about the RISC project and the green paper, please visit the project's webpage http://www.risc-project.eu