You are here
Intersect’s NSW-based supercomputer 'Orange' replaced McLaren in early 2013, continuing a tradition of naming for great Australian wine growing regions.
The SGI 34+ TFlop distributed memory cluster provide a greater than 25-fold increase of compute power and a fivefold increase of disk capacity on the previous system, McLaren.
February 2017: Important announcement for all Orange hpc.Time users...
We’re pleased to announce arrangements to transition supercomputing capacity from Orange.intersect.org.au to Raijin.nci.org.au under an agreement that has been in negotiation with our partners, the National Computational Infrastructure. Researchers will experience comparable capacity with greater scale, computational diversity, and, of course, performance. Intersect Australia's renown hpc.Time team expert advice, friendly service, training, support and flexibility will remain and be extended by the breadth and depth of NCI backing.
The FAQs below explain the rationale, arrangement, impacts and next steps as we work through this transition together.
Intersect and NCI have been in planning for some months and, with your understanding, we are confident of a rapid and successful migration. If you have any concerns or questions, please email email@example.com.
Frequently Asked Questions
1. Why is this happening?
We communicated in December last year that Orange.intersect.org.au was nearing the end of its working life and plans were being considered to transition to higher performing replacement technology early this year via ARC LIEF grant LE160100002, a successful grant application coordinated by Intersect with lead CI The University of Sydney.
The original intent of the LIEF grant was to make a capital investment in new generation HPC hardware, essentially updating Orange with current technology. It became clear during planning consultation in March 2016 that the consensus CI position was that aggregation at this quantum of investment scale is less relevant today and that greater computing leverage would be possible through pooling with NCI. This thinking was validated in subsequent consultation with Member DVCsR and CIOs.
Once NCI capacity was confirmed via the NCRIS Agility Funding announcements in late August, Intersect and NCI broadly agreed terms and the University of Sydney sought ARC consent to vary LE160100002 in September. In recent days, the ARC made a final determination to allow this to happen. Consequently, we are now executing a transition plan on an accelerated timescale to ensure all research projects and users are migrated with as little disruption as possible.
2. What about my capacity allocation?
Due to the delay in receiving ARC consent, we opted to defer the annual Resource Allocation Round in December, and all current allocations have been maintained to date consistent with membership agreements and commercial arrangements. The status quo will be migrated across to Raijin, and once Orange is fully decommissioned post-March we will revisit the question of ongoing resource allocation.
3. What will happen during this transition?
We will continue to operate Orange until 20 March 2017. During this transitional period Intersect and NCI will create user accounts and projects on Raijin. In order to drain the job queues on Orange there will be a cutoff date for job submissions. Intersect will copy all project data to Raijin and users resource allocation will be determined once users start to use their projects. Projects on Orange will be made read-only prior to the final closure of the machine and we will notify users when this is going to happen. After the final closure of Orange it will no longer be accessible to anybody, all storage will be erased and the infrastructure will be decommissioned.
4. What happens next for Orange users?
Dr. Joachim Mai's hpc.Time team will work with NCI to setup all users on Raijin. Project codes and usernames will be the same as on Orange but the password will differ. If your contact details are not current you’ll need to update them at https://my.nci.org.au/mancini - please get in touch with us if you can’t remember your Mancini credentials. New Raijin users will receive a notification with their new login credentials.
Our team will copy your project data to Raijin and will notify you when this is completed. If you need additional resources for the remainder of the quarter (until 31st March) please contact our team via firstname.lastname@example.org.
5. What's the timetable?
Standard job queue (250h walltime) closes for new jobs: 3rd March 2017
Last jobs on Orange completed: 15th March 2017
Projects will be made read only: 16th March 2017
Users can copy latest results back to their local machines or Raijin until: 18th March 2017 (this date may be subject to change depending on progress)
Orange machine closed and shut down: 20th March 2017
Orange officially decommissioned: 31st March 2017
6. What do I do next?
Please don’t submit jobs after the queue closure dates listed, as they either will not run or will necessarily have to be terminated.
If you have an existing account on Raijin we are working on providing additional SUs for the transition period and we will be in touch.
Users without accounts on Raijin will need to ensure that a correct mobile phone is associated in https://my.nci.org.au/mancini to allow passwords to be sent by SMS text message. We will get in touch to verify your account is operational.
You will be notified when your Raijin project is setup and available and data copied.
Getting started information about Raijin for Orange users is available here.
Detailed technical documentation about Raijin is available at NCI.
Orange features 100 cluster nodes with 1660 cores powered by the Intel® Xeon® E5-2600 processor series. It also includes 200TB local scratch disk space, 101TB of usable shared storage delivering 30+ TFlops.
The SGI HPC cluster is comprised of 13 large compute nodes each with dual Intel Xeon E5-2600 8-core processors, 256 GB memory, and 1x2TB SATA drives. System software provided includes SGI Management Center, SGI Performance Suite, PBS Pro Scheduler and SUSE® Linux Enterprise Server operating system. In addition, there are 90 small compute nodes each with dual Intel Xeon E5-2600 8-core processors, 64GB memory, 1x1TB SATA drives. The clusters are connected with QDR InfiniBand® Non-blocking Interconnect technology.
Dual administration nodes and a system console are also provided. Storage capabilities consist of an SGI NAS Storage Server with Panasas® ActiveStor™ 12 (high performance parallel file system) delivering 57TB usable storage.
Orange is funded through the Australian Research Council's Linkage Infrastructure, Equipment and Facilities (LIEF) scheme, (project number LE110100143). The LIEF grant, led by the University of Sydney’s Professor Leo Radom is supplemented by investments from the University of Sydney, UNSW, UTS, Macquarie University, the University of Newcastle, the University of Wollongong, Southern Cross University and the University of New England. The combined value of the capital investment is greater than $1million.
Intersect provides the on-going hosting facilities, management and support of HPC systems on behalf of the consortium of NSW universities. A rigorous procurement process was led by the University of Sydney.