Big-Data Science: Infrastructure Impact

Inder Monga, Prabhat Berkeley
2018 Proceedings of the Indian National Science Academy  
institutions, researchers, students, industry and academia. This is the only way that a nation can maximize the research capabilities of its citizens while maximizing the use of its investments in computer, storage, network and experimental infrastructure. This chapter introduces infrastructure requirements of High-Performance Computing and Networking with examples drawn from NERSC and ESnet, two large Department of Energy facilities at Lawrence Berkeley National Laboratory, CA, USA, that
more » ... ify some of the qualities needed for future Research & Education infrastructure. Most scalable Deep-learning Implementation National Energy Research Scientific Computing Center (NERSC) reported in their communication dated 28 August 2017, that a collaborative effort between Intel, NERSC and Stanford has delivered the first 15-petaflops deep learning software running on HPC platforms and is, according to the authors of the paper (and to the best of their knowledge), currently the most scalable deep-learning implementation in the world. The work described in the paper, Deep Learning at 15PF: Supervised and Semi-Supervised Classification for Scientific Data (https://, reported that a Cray XC40 system with a configuration of 9,600 self-hosted 1.4GHz Intel Xeon Phi Processor 7250 based nodes achieved a peak rate between 11.73 and 15.07 petaflops (single-precision) and an average sustained performance of 11.41 to 13.47 petaflops when training on physics and
doi:10.16943/ptinsa/2018/49338 fatcat:u6aogkxehzatvmjvcunjp3vn64