2017 HPC Summer School
Supporting over 4,000 researchers using the facility, NCI is committed to training new users to ensure research groups gain the maximum amount of value from their allocated time on the supercomputer, Raijin.
In February 2017, a five day HPC Summer School will run on campus at The Australian National University.
The NCI HPC Summer School is designed as a springboard for postgraduate research students and experienced research staff who wish to learn more about Raijin and high-performance computing.
Days One, Two and Three of the program will cover key topics, including:
– overview of NCI resources and services
– Linux essentials
– how to use Raijin, including the modules system, and PBS Pro job scripting
– parallel programming using MPI/OpenMPI
– application performance analysis and optimisation on Raijin
Days Four and Five will focus on advanced scientific application development using accelerators: GPUs and Knight’s Landing.
NVIDIA staff will provide an introduction to the fundamentals of GPUs and CUDA (the parallel language for GPUs). This is not a programming workshop; no formal programming experience is required. You only need to have a working knowledge of scripting or how your code runs, and a desire to learn.
They will explore one of the hottest applications of GPUs today: Deep Learning, as well as providing an overview of the latest innovations that are driving the hardware and software platforms as well as the Deep Learning SDK.
Research students who complete the program will have road-tested experience on Raijin, which can be applied to their own research. Those with prior experience on Raijin are also encouraged to attend the modules/workshops on advanced topics, e.g. performance and accelerators, which can be leveraged to improve their research throughput.
All current and future NCI users are encouraged to attend. The Summer School will be a great opportunity to meet and work closely with NCI staff.
NOTICE: Attendees are required to bring their own laptops.
Monday 6 February
Introduction to NCI and Raijin
Introduction to Linux
Tuesday 7 February
Parallel Programming (MPI/OpenMPI)
Wednesday 8 February
Introduction and Advanced Python
Thursday 9 February
GPU Workshop, hosted by NVIDIA
– Intro to GPU programming
– CUDA libraries and profiling
– Deep learning on GPU’s
– GPU-Accelerated Python
Friday 10 February
Participants are welcome to register for individual days or for the entire Summer School. Costs for one day only attendance are $60, and the five days is $250. Full registration includes the reception on Thursday evening, additional tickets can be purchased for $30.
Participants will be required to arrange their own transport and accommodation in Canberra. University House is a hotel on campus at ANU, and the venue is close to numerous hotels in New Acton and Canberra CBD.
Event details and tickets available on Eventbrite