OpenPOWER Application Optimization meet-up summary at Indian Institute of Science in Bangalore, India

Prof Gopinath at IISc, India in collaboration with OpenPOWER Academia hosted 3 hours OpenPOWER Application Optimisation meet-up where good set of researchers , students and industry folks participated . 

Here is the brief summary about the meetup 

Prof. Dirik from Julich Research Lab presented IBM Power9 processor, Nvidia V100 GPU, and IBM AC922 node hardware architecture. The interesting parts of the talk are performance analysis (bandwidth latency performance model) and the use case of solving dimensional Poisson equation. Another important point is how AC 922 node can be used as a building block for a super computer.

Dr Archana Ravindar from IBM talked on performance tuning using compilers is one of the key take away of the meet-up. It was mostly hands-on session and she demonstrated the performance improvement (speed-up) with various GCC compiler options including software pre fetching, OpenMP parallelization, and GPU acceleration.

3. Dr. Prathu from nVidia's presentation was more an interaction to GPU and various features . 

Video presentations also available in OpenPOWER AI Virtual University channel

Overall the meetup did involve good discussions around Power 9 Processor features , nVidia GPU capabilities and IBM Power 9 system AC922 system's capabilities.

I would like to thank Dr. Gopinath who supported the meet-up , Dr. Pleiter , Dr Archana and Dr Prathu who shared their experiences working on Power 9 systems as part of the meet-up