Date

Sage2 will be participating ing HP-C/DA 2021 Workshop on the In Situ Co-Execution of High-Performance Computing & Data Analysis

HP-C/DA takes place Friday, July 2, 2021, 2PM CEST

You are all invited to join HP-C/DA 2021!
The workshop on In Situ Co-Execution of High-Performance Computing & Data Analysis.
**Friday, July 2, 2021, 2PM CEST**, a workshop of ISC High Performance.

https://hpcda.github.io/

HP-C/DA focuses on the latest advances and challenges for the co-execution of HPC & HPDA workloads on supercomputers. It brings together all stakeholders concerned by the in situ co-execution of HPC and HPDA workloads:

  • computer-scientists interested in high-performance applications   modularization and coupling,
  • designers and developers of services and tools for HPDA on supercomputer   architectures,
  • application developers interested in the efficient integration of HPDA in   their HPC  application,
  • application scientists concerned with using novel and innovative HPC &   HPDA couplings at scale.

HP-C/DA  will act as a discussion forum for all parties involved in order to identify the best research directions and gather requirements from the user
community.

Exascale computing offers the promise of supporting disruptive numerical experiments needed to address the pressing scientific, industrial and societal challenges of the 21st century, such as clean energy, health and climate change. ExaFlop/s supercomputers will provide the compute capabilities to support simulations of yet unseen precision, generating an unprecedented quantity and quality of data. Only the latest advances in automated data analytics based on machine-learning or statistical analysis will make it possible to get the most knowledge out of the data generated at this scale.
 

To reach and go beyond Exascale, it is critical to consider the numerical experiment as a whole, encompassing both the simulation (high-performance computing, HPC) and data-analytics (high-performance data-analytics, HPDA) aspects. Optimized simulation codes, at the core of the numerical experiments, need to be augmented with advanced data analytics in tight coupling patterns so as to overcome the widening performance gap between compute and I/O, and leverage new deep memory hierarchies. Once this technical barrier overcome, innovative numerical patterns become accessible where the outcome of analytics is used to steer the simulation. These new patterns have the potential to greatly increase the scientific return of investment for numerical experiments at all scales from lab-scale clusters to the largest supercomputers.