The Santos Dumont Petaflop Machine and the Brazilian HPC Research Challenges

Carla Osthoff
Laboratório Nacional de Computação Científica Petrópolis, Brazil

Abstract: On July 3, arrived at the headquarters of the National Laboratory for Scientific Computing (LNCC) a supercomputer, which was named after Santos Dumont, or SDumont briefly. SDumont own processing capacity of 1.1 Petaflop/s, with a hybrid configuration of compute nodes, with regard to the available parallel processing architecture. The SDumont has a total of 18,144 CPU cores, distributed in 756 compute nodes, which are comprised mostly only on the CPUs multi-core architecture. The acquisition of this new computer from the French company ATOS/BULL is part of the Strategic Program for Software and Information Technology Services from the Ministry of Science, Technology and Innovation, by the Brazil/France High Performance Computing agreement. The supercomputer is expected to further boost the research sector, causing major impact on scientific development and international collaborations.

About the speaker: Carla Osthoff is a researcher at National Laboratory for Scientific Computing (LNCC-Brazil). Her main research areas are: High performance Computing, Cluster Computing, CPU/GPU/MIC Hybrid Parallel Computing, High Performance Scientific Computing Applications and Parallel Programming Models. She has worked in several research projects, among which we found: Petaflop01, Aplicações do Processamento de Alto Desempenho com arquiteturas de processamento híbridas and Atmosfera Massiva-Impacto de Novas Arquiteturas de Processadores em Modelos Atmosféricos de Produção II.

Profr. Osthoff received her BSc in Electronics Engineering from Pontifícia Universidade Católica do Rio de Janeiro, PUC-Rio, Brazil, a MSc and a PhD in Computer Science from Universidade Federal do Rio de Janeiro, UFRJ, Brazil. She was part of the hardware engineer research group at the Brazilian Center for Physics Research (CBPF/ Brazil), participating in the development of multiprocessors parallel systems with Fermilab/ USA from 1986 to 1996.

She is currently the coordinator from LNCC High Performance Computing Center, from LNCC GPU Computing Research and LNCC GPU Teaching Center, several LNCC High Performance Computing Research collaboration projects and Santos Dumont Supercomputer Technical-Scientific Advisory Committee member. She has published more than 40 papers between journals and congress proceedings, she has published 3 book chapters and edited one book on HPC.
Back Keynotes