Pawsey Uptake Project – Expression of Interest Call
Pawsey Uptake Projects are collaborative projects to maximise the impact of research using Pawsey’s services. The key goals for uptake projects include:
- Increasing the scale and scope of scientific research using Pawsey infrastructure.
- Improving the efficiency or performance of scientific workloads using Pawsey supercomputers, including computational, visualisation, and data management workflows.
- Adoption of emerging technologies such as machine learning, artificial intelligence, or quantum computing.
- Encouraging uptake from a wide range of scientific domains, including bioinformatics.
This expression of interest call for Pawsey Uptake Projects is open to Australian-based research communities. The project leader must be employed by an Australian university, government or research institution. Successful projects receive up to 0.20 FTE of a dedicated Pawsey staff over six months to work in collaboration with the research team aiming to improve their usage of the Pawsey infrastructure. Depending on the nature of the project, this time may be distributed across expertise from multiple Pawsey staff if necessary. The distribution and total effort provided will be determined by Pawsey, based on the proposed project. Although this is not an allocation call, new projects may be granted a Preparatory Access allocation for benchmarking and development if needed.
The scope of an individual project may include, but is not limited to:
- Feasibility studies involving profiling and benchmarking scientific software to identify scope for future improvement.
- Improving performance by migrating interpreted scripts (such as MATLAB, R, Python) to compiled languages (such as C/C++ or Fortran).
- Optimisation of scientific software using code refactoring to improve vectorisation, use of high performance numerical libraries and/or use of GPUs.
- Parallelisation and acceleration of scientific software using MPI, OpenMP, ROCm, and/or HIP.
- Improving code performance by identifying and correcting performance bottlenecks related to data transfer and/or communication.
- Improve the usability of the application by adding support for outputting data using common interchange formats (such as netCDF or HDF5).
- Optimising workflows for more efficient queue utilisation, improved job scripts, or use of containers for complex software stacks (such as using Nextflow or Snakemake).
- Managing complex data ingestion, analysis, and/or sharing workflows that span multiple Pawsey storage systems.
- Investigating the configuration or adoption of a proof-of-concept for new features, APIs, services, or protocols.
- Developing workflows to effectively visualise large-scale/complex datasets, including in-situ visualisation using remote visualisation.
- Developing scripting to generate visualisations of high-quality images and animation using remote rendering.
- Visualisation of dataset using the remote VR service.
- Translation of classical algorithms to quantum computing algorithms and development of hybrid workflows for running on both quantum computing hardware (or simulators) and classical HPC hardware.
- Efficient use or incorporation of machine learning and artificial intelligence tools for your research.
Projects aiming to use Setonix’s power-efficient GPU architecture will be prioritised. Depending on the nature of the collaboration, there may also be opportunities for consulting and participation in hackathons and code sprints. The collaboration is expected to result in a project report, and preferably joint publications.
Please apply via the Pawsey Application Portal, using the Pawsey Uptake Project form. Note that the application portal uses your AAF institutional credentials for authentication rather than your Pawsey account. For assistance with your application or any other enquiries, please contact the Pawsey help desk via the User Support Portal or email at help@pawsey.org.au
Applications close Monday, 22nd of April 2024, End of Day, Anywhere on Earth (AoE).