“Faster, more efficient software,” “speedups and code profiling,” and “parallel programming” – these were the central takeaways from the two-week National Computational Infrastructure (NCI) HPC-AI Hackathon hosted in partnership with NVIDIA, and the OpenACC organisation from 24 October to 4 November. With attendees participating online and onsite on The Australian National University campus in Canberra, the HPC-AI Hackathon was a significant opportunity for teams from across the region to develop improvements to their codes to take advantage of the modern graphics processing units (GPUs) available to them.
“It has been an exciting and productive week of intensive coding experience for me and my students at NCI’s GPU Hackathon,” said Dr Kenneth Duru from The Australian National University. The group is developing a dynamic earthquake simulation code to understand seismic risks and hazards. “With help from mentors across NCI, NVIDIA, and ANU, we have made significant steps towards accelerating our research code WaveQLab3D by enabling GPU computations. Using a hybrid CPU-GPU implementation has been a long-held ambition for us.”
Modern programming has benefited from GPU accelerators for more than two decades, and now scientific applications are taking advantage of these performance gains. The recent boom in AI has made accelerators even more critical. Nine teams with more than 50 researchers from across Australia, Taiwan, India, Singapore, and the US worked together and with the NCI and NVIDIA experts to crack open their codes and make them more efficient. The science domains covered climate and weather science, geophysics, material science, fluid dynamics, aircraft modelling and computational chemistry. All teams created substantial improvements to their HPC or AI applications. NCI is proud to support cutting-edge research around Australia and the world.
This Hackathon is one of several high-impact international training events hosted at NCI this year. Through a range of bootcamps in 2022, we have been building our users’ capacity to make good use of our GPU resources. The precursor events, bootcamps in data science, distributed deep learning, NVIDIA CUDAⓇ Python, OpenACC and CUDA C/Fortran, provided hands-on learning and exercises for skills development. Participants learned how to profile their code, find the computational sweet spots, and improve their performance with different strategies and tools. Finally, during the hackathon they were able to apply what they learned to their own applications. This was the culminating opportunity for NCI users to get dedicated help and guidance from both NCI and NVIDIA mentors.
“This is the first time NCI has hosted such an intensive and user-focused hackathon. It was only possible to make it this big through our collaboration with NVIDIA and the OpenACC organisation through the Open Hackathons program. It is very satisfying to observe the achievements of the teams over the course of the two weeks,” said Dr Jingbo Wang, Training and User Services Manager at NCI.
“The engagement between mentors and research teams went extremely well. We filled their skills gaps by providing guidance on their applications, and they shared with us their background and experience so that we can better understand the education, training and technology needs of our users in the future. It was just an amazing event to be a part of.”
Bharat Kumar, GPU Advocate at the NVIDIA AI Technology Center said, “Open Hackathons is a platform where domain scientists learn high performance computing (HPC) by applying the concepts directly to their own research codes with guidance from mentors. This NCI Hackathon was a true global event with mentors coming from Australia, Taiwan, Singapore, and India and applications ranging from Fluid Dynamics to Climate, Materials Science, Computational Chemistry and many more.”
NCI’s ongoing training collaboration with NVIDIA has helped users to develop expertise with their codes directly on the Gadi system, working in real-time on the platform they use every day. The Hackathon has also provided an opportunity to develop the mentors’ skills in training, education and content delivery.
Team mentor and NCI Senior HPC Systems Specialist Dr Ben Menadue said, “Having everyone under the one roof allows for efficient collaboration and forthright discussion. The expertise of the NVIDIA mentors compliments the detailed system knowledge of the NCI staff, and the close interaction provided by the Hackathon environment helps to guide the teams in developing, debugging, and improving the performance of their applications.”
The NCI-NVIDIA HPC-AI Hackathon has set up researchers to perform bigger, better and more efficient calculations with the HPC resources available to them. Like for Dr Kenneth Duru’s earthquake studies, the Hackathon is helping build the skills of researchers and capacities of important scientific software. It has also laid the groundwork for future national and international training events. NCI is looking forward to hosting further GPU training events for our users and international computational researchers in coming years.
The teams participating in the Hackathon were:
- Hypersonics Team (UNSW Canberra)
- CLEX Team (UNSW/CLEX)
- Seis Team (ANU)
- WaveQLab3D Team (ANU)
- Rxn-net Team (UNSW)
- LTRAC-GO Team (Monash University)
- Hyperhacks Team (UQ)
- EXESS Team (ANU)
- Water Research Centre Team (UNSW/University of Melbourne)