When you submit your application through our web site, it will go to our technical support team for review. In general, our review process for new projects is as follows:
- Initial check: We see whether your proposal describes a legitimate research programme and whether your research programme will need some kind of advanced research computing capability (which may or may not be high-performance computing). We also check whether your project team is all assembled and has the skills needed to start using our systems (for example, basic familiarity with the Linux command line).
If you are are member of NIWA we will also confirm with NIWA's institutional point of contact that you have followed the NIWA internal documentation for gaining access to the HPC's. Note, you will only be able to access the NIWA internal documentation if you are currently behind the NIWA VPN.
- Software check: One of our technical experts looks at the software you say you want to use and determines whether it can run on any of our systems and whether you are likely to be legally allowed to run the software on NeSI. This check is intended to cover both compatibility and licensing matters, as well as whether you are able and willing to compile or install the software yourself if necessary.
- Support check: Some research programmes may have very demanding support needs. We will consider whether we are able to offer the kind and amount of support your team is likely to need to progress your research if we approve it. This check is especially important if we think you are likely to want or need to change someone else's code. We may consult with our scientific programmers at this point, and find out whether your project is likely to be eligible for our consultancy service.
- Disk space check: We decide how much disk space your project is likely to need in the persistent storage (project directory) and scratch storage (nobackup directory). We may unfortunately have to reject (or negotiate for less storage) if your disk space needs would interfere with our ability to provide good service to other research teams.
- Facility: Based on the information in your application, we decide whether your workflow is best suited for Mahuika, Māui or both, and also whether your project would benefit from an allocation of GPU hours or access to ancillary nodes or virtual labs.
- Decision and notification: If we approve an initial allocation for your project, we will typically award the project an allocation of CPU core hours (and if necessary GPU core hours) from either the Proposal Development class or the Institutional class. In any case, we will send you an email telling you about our decision.
Our review process for requests for new allocations on existing projects is simpler:
- Eligibility check: We look at the information you have given us (and may ask you more questions) to find out which of our regular allocation classes (Merit, Postgraduate or Institutional) this research programme is eligible to receive. Your research programme may be eligible for more than one allocation class.
- Amount and duration: We will choose how many CPU core hours you are likely to need based on what kind of allocation you most recently received and your usage history. We may suggest an allocation size (i.e. number of CPU core hours, node hours, and/or GPU hours) and a duration of up to 12 months, and give you a chance to provide feedback if you think our suggested allocation would not meet your needs.
- Choice of Class and Contention: We will choose from which class to award your allocation, based on your research programme's eligibility for the different classes and whether your proposed allocation would exceed any class-based allocation limits. We may change this choice depending on which classes, if any, are under contention.
- Approval: If we decide that your project should be considered for an Institutional allocation, the request may need to be approved by a representative of the project's host institution, which is the institution where the project owner works or studies.
- Decision and notification: We will send you an email telling you about our decision.
From time to time we may have to decline requests for allocations of computing resources. If we can't grant your research programme an allocation because of contention or because the project's host institution has disallowed the allocation, this is not to be taken as a judgement on the merit of your research topic or the quality of your team's work.