J. Laberta, BSC (ES)

Abstract: Squeezing performance out of complex and continuously changing architectures is difficult. Refactoring huge codes for next-generation machines is OK… But only until a next-next generation comes. And continuous refactoring is a dangerous loop that domain scientists rightly want to avoid. Trying to foresee the weather forecasting community’s curiosity, I try to answer the questions, “Which will be the next HPC system? What impact will have on our codes? Which is the status of future computing systems in Europe?”.
The talk guided the audience through the European Processor Initiative (EPI) project and in particular its RISC-V vector accelerator targeting HPC. Since the audience should be familiar with HPC applications, the idea was to start from the performance analysis of weather forecasting HPC codes to show how to get insights and architectural implications that can influence the design of future HPC systems.