Currently, Nimbus consists of AMD Opteron Central Processing Units (CPUs) making up 3000 cores and 288 terabytes of storage.
The expansion includes 6 x HPE SX40 nodes which each contain 2 x NVIDIA Tesla V100 16GB GPUs. These bad boys are built to accelerate Artificial Intelligence, HPC and graphics. Powered by NVIDIA Volta architecture, they offer the performance of up to 100 CPUs in a single GPU.
With the GPUs currently being installed, the Pawsey cloud team have opened a Call for Early Adopters. Pawsey are seeking researchers to provide user time on the system to aid Pawsey in validating and optimising the service in preparation for full production use.
The Early Adopter program will run until August 2018, however this schedule may vary depending on the duration of commissioning and acceptance testing.
This expansion is going to provide a new set of functionalities for researchers. If it sounds familiar, it is because Pawsey are now playing in the same leagues as Google, Amazon, IBM and Microsoft who have recently announced their cloud expansion using the same NVIDIA GPUs.
One of the differences between these key cloud players and Pawsey are that the services offered at the Centre are FREE to Australian researchers. That means, any project affiliated with an Australian university, research body or government facility can use the service at no cost.
To apply as an early adopter, the Nimbus GPU (Early Adopter) application form can be completed.
The Early Adopter period will remain open until the system is fully commissioned.
The Nimbus GPU expansion was announced at the launch of Data Science Week 2018 in Perth, Western Australia.
Data Science Week Grab is a series of events showcasing how data science is applied to areas such as supercomputing, machine learning, computation and analytics. With more than 15 events scheduled for the week, visit the website www.datascienceweek.org or follow the hashtag #DSW18!
Source: Pawsey Supercomputing Centre