ANU HPC Team

To design, optimise, and deploy high-performance computing systems that maximise efficiency, scalability, and innovation, empowering our team to compete at the highest level.

ANU HPC Mission Statement

About Us

We are a passionate HPC team focused on solving complex computational problems, developing advanced technical skills, and competing at the highest level.

We are a team of ANU students driven to compete at the highest level in the Student Cluster Competition. Our goal is to design, build, and optimise a high-performance computing cluster that delivers maximum performance under strict power and cost constraints. Through rigorous benchmarking, system tuning, and teamwork, we aim to push our hardware and skills to their limits while representing ANU on the global HPC stage.

Alongside our competition efforts, we are developing a GPU-accelerated version of the IQ-TREE phylogenetic software. This project focuses on porting computationally intensive components to run efficiently on modern GPU architectures, significantly reducing runtime for large-scale evolutionary analyses. By combining parallel programming techniques with HPC optimisation strategies, we aim to make cutting-edge bioinformatics tools faster and more accessible for researchers.

Duck

Our Hardware

Our systems are engineered to balance performance, power efficiency, and scalability, enabling us to tackle demanding HPC workloads and competition benchmarks.

XENON GPU Cluster

16× A100 40GB SXM4 GPUs  ·  8× AMD EPYC 7313 CPUs
  • High-end GPU acceleration for AI & HPC workloads
  • Hybrid CPU-GPU workloads leveraging AMD EPYC cores alongside A100s
  • Optimised for CUDA, OpenMP, and hybrid MPI+GPU workloads
  • Used for IQ-TREE GPU port and parallel compute tasks

Raijin Test Cluster

12× CPU Compute Nodes
  • Multi-node distributed system for MPI benchmarking
  • Used for HPL tuning and SCC workload simulation
  • Supports hybrid MPI + OpenMP configurations
  • Ideal for scaling tests and network performance analysis

NCI Gadi Supercomputer

National HPC Facility
  • Access to thousands of CPU cores and GPU nodes
  • Used for large-scale simulations and validation runs
  • PBS Pro workload scheduling environment
  • Supports production-level HPC workflows

Pawsey Setonix Supercomputer

Pawsey Supercomputing Research Centre
  • One of the world's most energy-efficient supercomputers
  • Access to large-scale CPU and GPU partitions
  • SLURM-based scheduling across thousands of nodes
  • Supports GPU-accelerated and memory-intensive workloads

Our Sponsors

We are proud to be supported by academic and industry partners who help us develop, build, and optimise high-performance computing systems.