Glossary
- Cristian Di Pietrantonio
- Alexis Espinosa
| Term/acronym | What it stands for... | What it means... |
|---|---|---|
| AAF | Australian Access Federation | Established in 2009, the Australian Access Federation (AAF) is Australia’s leading identity broker, enabling access to online resources for the Education and Research sector |
| AARNet | Australian Academic and Research Network | https://www.aarnet.edu.au/about-us Ultra high-speed internet and collaboration services only for research and education. |
| allocation | Budgeted compute time on Pawsey supercomputers that is awarded to researchers and commercial users. Allocations are measured in service units (SUs). | |
| AMD | Advanced Micro Devices, Inc | Vendor of CPUs (EPYC™) and GPUs (Instinct™) used in the Setonix system. |
| API | Application Programming Interface | System design to respond to requests of a specific format. These are locked down to IP ranges that may receive a response. |
| ARDC | Australian Research Data Commons | The ARDC enables the Australian research community and industry access to nationally significant, data intensive digital research infrastructure, platforms, skills and collections of high quality data. |
| ARRC | Australian Resources Research Centre | This is most typically used to refer to the building in which many of the Pawsey staff are based, as distinct from the building that actually houses the supercomputers operated by the Pawsey Supercomputing Research Centre. |
| ASKAP | Australian Square Kilometre Array Pathfinder | https://www.atnf.csiro.au/projects/askap/index.html |
| bucket | A bucket is a named container for objects stored in an object storage system such as Acacia. | |
| data mover node | A node of a supercomputer dedicated to executing large file transfers. | |
| exclusive node usage | Accounting model that charged for whole nodes irrespective of whether they had been fully used. (This model was used on earlier Pawsey systems before Setonix.) | |
| FAIR | Findable, Accessible, Interoperable, Reusable | The FAIR Data Principles (Findable, Accessible, Interoperable, Reusable) were drafted at a Lorentz Center workshop in Leiden in the Netherlands in 2015. There is a push to make research data more "FAIR" |
| HPC | High Performance Computing | Term for all things 'super computer'; enables computationally-intensive research. |
| GPU | Graphics Processing Unit | A processor comprising many specialized cores. GPUs can deliver massive performance improvements for tasks that can be processed in parallel across many cores. Originally designed to move graphics-intensive tasks off the CPU, GPUs are ideal for many SC applications including manipulating very large data sets. |
| HPE | Hewlett Packard Enterprise | Vendor for the Setonix system (Cray EX architecture) to replace Magnus and Galaxy. |
| job arrays | A SLURM feature whereby the same batch script is submitted multiple times, but each instance is assigned a different identifier that can be used to differentiate the execution path within the script. | |
| job chaining | Coding a series of batch jobs so that the end of each job script submits the next job in the chain. | |
| job packing | Running multiple independent programs within a single slurm Job for better resource utilisation and to reduce the number of jobs submitted to the scheduler. | |
| LTS | Long Term Storage | New Pawsey storage system comprising both LTS Online (disk) and LTS Offline (tape) subsystems. This is a working name until it (or they) receive naming. Part of the Pawsey Capital Refresh project. |
| Lustre | A high-performance, highly scalable shared filesystem for Linux clusters. Each Lustre filesystem is composed of multiple servers, with one or more metadata targets (MDT) and multiple object storage targets (OSTs). The Lustre software presents the OSTs as a single unified filesystem. For more definitions, see: Acronyms in the Lustre wiki. | |
| MDS | Metadata Server | Service nodes that manage all metadata operations, such as assigning and tracking the names and storage locations of directories and files on the OSTs. There is at least one MDS containing one or more MDTs per Lustre filesystem. |
| MDT | Metadata Target | A storage device where the metadata (name, ownership, permissions and file type) are stored. There is at least 1 MDT per Lustre filesystem. |
| MWA | Murchison Widefield Array | The Murchison Widefield Array (MWA) is a low-frequency radio telescope operating between 80 and 300 MHz. It is located at the Murchison Radio-astronomy Observatory (MRO) in Western Australia, the planned site of the future Square Kilometre Array (SKA) lowband telescope, and is one of three telescopes designated as a Precursor for the SKA. |
| NCI | National Computation Infrastructure Australia | NCI Australia is a highly-integrated, high-performance research computing environment. NCI is built to deliver on national priorities and research excellence from across the scientific disciplines. |
| NCMAS | National Computational Merit Allocation Scheme | NCMAS) is Australia’s premier meritorious allocation scheme, spanning both national peak facilities and specialised compute facilities across the nation. NCMAS is open to the Australian research community, providing significant amounts of compute time for meritorious research projects. |
| NeCTAR | National eResearch Collaboration Tools and Resources project | https://ardc.edu.au/services/nectar-research-cloud/ Nectar provides an online infrastructure that supports researchers to connect and collaborate with colleagues in Australia and around the world. |
| NERSC | National Energy Research Scientific Computing Centre | http://www.nersc.gov/ The National Energy Research Scientific Computing Center (NERSC) is the primary scientific computing facility for the Office of Science in the U.S. Department of Energy. |
| NVMe | Non-Volatile Memory Express | Fast storage interface for SSDs. NVMe is a communications standard that operates across the PCIe (PCI Express) bus. This allows SSDs "to act more like the fast memory that they are, rather than the hard disks they imitate". |
| object storage system | A non-hierarchical storage architecture where each unit of storage is an object. An object consists of data, such as a file, part of a file, or a sequence of bytes; metadata; and a globally unique identifier. Related objects are stored in a bucket. | |
| Origin | The Pawsey user and resource management system. | |
| OSS | Object Storage Server | Service nodes that run the Lustre software stack, provide the actual I/O service and network request handling for the OSTs, and coordinate file locking with the MDS. Each OSS can contain multiple OSTs. |
| OSTs | Object Storage Targets | Individual small filesystems that are presented by Lustre as a single unified filesystem. |
| PaCER | Pawsey Centre for Extreme scale Readiness | To prepare Australian computational researchers for the next era of supercomputing. The focus of the PaCER program is on both extreme scale research (algorithms design, code optimisation, application and workflow readiness) and using the computational infrastructure to facilitate research for producing world-class scientific outcomes. |
| PPS | Pawsey Partner Scheme | |
| PRACE | Partnership for Advance Computing in Europe | An international not-for-profit association comprising 26 member countries whose representative organisations create a pan-European supercomputing infrastructure, providing access to computing and data management resources and services for large-scale scientific and engineering applications at the highest performance level. |
| proportional node usage | Accounting model that charges only for the portion of a node that is requested. (This is the model used for Setonix.) | |
| PPI | Pawsey Partner institutions | Pawsey partner institutions are CSIRO, Curtin University, Edith Cowan University (ECU), Murdoch University and The University of Western Australia (UWA). |
| pseudo folders | In an object storage system, the forward slash ( / ) can be used as a separator in object names to imply a subfolder structure. | |
| PSRC | Pawsey Supercomputing Research Centre | The PSRC is a joint venture between CSIRO, Curtin, ECU, Murdoch and UWA. Its primary function is to accelerate scientific research for the benefit of the nation. It has expertise in supercomputing, data, cloud services and visualisation. Research domains include astronomy, life sciences, medicine, energy, resources and AI. |
| reproducibility | Creating job scripts and their runtime environment without local or user dependencies. The intention is that Pawsey staff or colleagues can run the same job and get the same results. | |
| reservation system | A system whereby specific nodes can be reserved at a specific time. Reservations are best suited for activities such as interactive debugging on many nodes that should run at a certain time of day. | |
| ROCm | Radeon Open Compute platform | https://github.com/RadeonOpenCompute/ROCm ROCm is focused on using AMD GPUs to accelerate computational tasks such as machine learning, engineering workloads, and scientific computing. |
| SC | Supercomputing or supercomputer | As in "SC architecture". |
| Service Desk | This application allows the lodgement, management and reporting on tickets which Pawsey manage over time. It allows the control of tasks through defined workflows and the handling of Changes in systems. | |
| Setonix | A new HPE/Cray EX supercomputing system commissioned to replace Magnus and Galaxy from 2021. | |
| shared access | Each compute node of Setonix can run multiple jobs in parallel, submitted by a single user or many, from any project. Sometimes this configuration is called shared access. | |
| Singularity | FOSS program that performs operating-system-level virtualization (containerization). Singularity brings containers and reproducibility to scientific computing and the high-performance computing (HPC) world. | |
| SKA | Square Kilometre Array | A very large radio telescope, aspiring to be the equivalent of a single dish with a square kilometre of collecting area |
| SLURM | Simple Linux Utility for Resource Management | A FOSS job scheduler for Linux kernels, used on many supercomputers and computer clusters. |
| SMT | Simultaneous Multithreading | A feature of modern CPU cores that allows them to execute two streams of instructions concurrently. |
| Spack | "Spack is a package manager for supercomputers, Linux, and macOS. It makes installing scientific software easy." | |
| SU | Service Unit | Allocations of compute time on Pawsey supercomputers are measured in service units (SUs). How SUs are defined and consumed varies across Pawsey systems. |
| system-specific pages | ||
| VM | Virtual Machine | A VM is an emulation of physical computer in the 'cloud'. Pawsey's cloud system is Nimbus (https://pawseysc.github.io/using-nimbus/) |