The Phoenix High Performance Computing (HPC) service enables researchers to leverage the expertise and increased computational power of a central on premise HPC facility called Phoenix.
Phoenix allows access to large clusters of CPUs and GPUs in order to solve problems that require large-scale computation. Phoenix HPC can handle large sets of data or especially complex tasks so researchers can perform big data analysis, complex modelling and simulations, to solve large scale research problems quickly and efficiently.
Who can use Phoenix?
Phoenix is largely utilised by Researchers for approved University of Adelaide research data processing.
Higher Degree Research Students can utilise Phoenix under the approval and supervision of their supervisor.
Teaching Purposes (students)
There may be situations where a student may need to use Phoenix to process large data sets for their university learning. This would need to be approved by the Division of Research and Innovation team prior to access being granted.
Who is responsible for Phoenix?
There are multiple layers of responsibility for Phoenix.
Users of Phoenix
When you agree to use Phoenix you will need ensure your use of it is for University of Adelaide purposes, unless you have prior approval from the Division of Research and Innovation allocation team. All users of Phoenix are required to clean and maintain their own data that is analysed. Phoenix has not been designed as a long term storage solution.
All users are required to have completed the introductory course in MyUni for Phoenix prior to accessing Phoenix. If you are new to HPC there are multiple learning opportunities outside of the University that can assist with preparing data to be processed.
For governance, approvals, investment in or feedback on Phoenix, please contact the Division of Research and Innovation.
If you are having issues with the technology infrastructure of Phoenix, eg if it is not working, you should contact the Information Technology and Digital Services - Service Desk (see below for contact details).
Features and benefits
Phoenix HPC allows University researchers to improve time frames for analysis and processing of big data or solving problems where there is increased complexity in models and simulations.
Using Phoenix assists researchers to improve research outcomes as it:
- Accelerates problem solving using higher capacity computing solutions, reducing run time by orders of magnitude
- Facilitates the exploration of bigger grids and parameters, reduces element size and can handle computations of increased complexity
- Allows simulations of many more scenarios and variables than possible on an office PC
- Provides larger amounts of memory available for your research.
Phoenix gives researchers a platform that provides better performance with no financial outlay, allowing researchers to leverages HPC research expertise using state-of-the-art equipment.
There are multiple research computing options available to researchers at the University. Before registering for Phoenix you may want to look at the High Performance Compute Options page to assist you in selecting the most appropriate compute option for your research.
Phoenix supercomputer hardware is a Lenovo NeXtScale system consisting of 260 nodesover 900 CPUs and GPUs. It is available on premise and can be accessed remotely.
Key System Features:
- Intel X86-64 Haswell and Skylake CPUs with NVidia GPU providing performance acceleration
- Connects to the 2PB Lustre high performance file system
Scheduling your jobs on Phoenix:
Phoenix uses the SLURM scheduler to manage the compute workload. After a job is submitted to a queue (also referred to as a partition), a number of factors are used to determine the priority of the job and when it will run. The factors are:
- Fairshare factor
- Job size
- Job partition
After the job is assigned a priority the scheduler will determine when the job runs based on the availability of the requested resources (CPUs, GPUs, memory), the job's requested wall-time and its priority.
To find out more about the underlying architecture of the Phoenix Supercomputer visit the HPC wiki page (https://wiki.adelaide.edu.au/hpc/index.php/Main_Page).
How much does it cost?
A new allocation model for Phoenix is currently under development which will include a base allocation available to all researchers, additional allocations on a merit basis, as well as a direct investment model.
How can I get access to Phoenix?
To start using Phoenix, you first need to register for an account.
The lead time for obtaining an account on Phoenix is a maximum of a week. For HDR students this is subject to authorisation from your supervisor.
If you have submitted an application for access to Phoenix and have heard nothing after one week, or for any other support, please contact the service desk using the provided support contact information below.
All users of Phoenix must adhere to the conditions associated with using Phoenix. These conditions are detailed at Conditions of use. This includes acknowledging Phoenix in any publications where you have utilised Phoenix as part of your research and that you only use Phoenix for your approved university research.
If you've already registered for Phoenix and want to get started, follow the instructions on the Phoenix High Performance Computing (HPC) wiki page.
Email: ITDS Service Desk firstname.lastname@example.org
Tel: +61 8 8313 3000
Technical Support for major incidents is available 24/7
Other technical support is available Monday to Friday from 9am to 5pm
The Phoenix HPC Wiki page has really valuable information about Phoenix that you will want to become familiar with as you use the machine. This includes information about Phoenix, Self Help Guides, Advanced Support, Training, Application-specific Information and Help Services.
Introduction to Phoenix
Other HPC Training Options
Self Help Guides
the Phoenix High Performance Computing wiki page.
If you need help with understanding you HPC or compute needs or need some advice on the best possible technology solution to meet your needs please visit the Consultancy and Advice page (page under development).