Skip to content
Contact Support


Mahuika is a Cray CS400 cluster featuring Intel Xeon Broadwell nodes, FDR InfiniBand interconnect, and NVIDIA GPGPUs.

Mahuika is designed to provide a capacity, or high throughput, HPC resource that allows researchers to run many small (from one to a few hundred CPU cores) compute jobs simultaneously, and to conduct interactive data analysis. To support jobs that require large (up to 500GB) or huge (up to 4 TB) memory, or GPGPUs, and to provide virtual lab services, Mahuika has additional nodes optimised for this purpose.

The Mahuika login (or build) nodes, mahuika01 and mahuika02, provide access to GNU, Intel and Cray programming environments, including editors, compilers, linkers, and debugging tools. Typically, users will ssh to these nodes after logging onto the NeSI lander node.


  1. The Cray Programming Environment on Mahuika, differs from that on Māui.
  2. The /home, /nesi/project, and /nesi/nobackup filesystems are mounted on Mahuika.
  3. Read about how to compile and link code on Mahuika in section entitled: Compiling software on Mahuika.
  4. An extension to Mahuika with additional, upgraded resources is also available. see Milan Compute Nodes for details on access

Mahuika HPC Cluster (Cray CS400)

Login nodes

72 cores in 2× Broadwell (E5-2695v4, 2.1 GHz, dual socket 18 cores per socket) nodes

Compute nodes

8,136 cores in 226 × Broadwell (E5-2695v4, 2.1 GHz, dual socket 18 cores per socket) nodes;
7,552 cores in 64 HPE Apollo 2000 XL225n nodes (AMD EPYC Milan 7713) the Milan partition

Compute nodes (reserved for NeSI Cloud)

288 cores in 8 × Broadwell (E5-2695v4, 2.1 GHz, dual socket 18 cores per socket) nodes


9 NVIDIA Tesla P100 PCIe 12GB cards (1 node with 1 GPU, 4 nodes with 2 GPUs)

8 NVIDIA A100 PCIe 40GB cards (4 nodes with 1 GPU, 2 nodes with 2 GPUs)

16 NVIDIA A100 HGX 80GB cards (4 nodes with 4 GPU each)


Enabled (accordingly, SLURM will see ~31,500 cores)

Theoretical Peak Performance

308.6 TFLOPs

Memory capacity per compute node

128 GB

Memory capacity per login (build) node

512 GB

Total System memory

84.0 TB


FDR (54.5Gb/s) InfiniBand to EDR (100Gb/s) Core fabric. 3.97:1 Fat-tree topology

Workload Manager

Slurm (Multi-Cluster)

Operating System

CentOS 7.4 & Rocky 8.5 on Milan

Storage (IBM ESS)

Scratch storage 4,412 TB (IBM Spectrum Scale, version 5.0). Total I/O bandwidth to disks is ~130 GB/s
Persistent storage 1,765 TB (IBM Spectrum Scale, version 5.0). Shared between Mahuika and Māui Total I/O bandwidth to disks is ~65 GB/s (i.e. the /home and /nesi/project filesystems)
Offline storage Of the order of 100 PB (compressed)

Scratch and persistent storage are accessible from Mahuika, as well as from Māui and the ancillary nodes. Offline storage will in due course be accessible indirectly, via a dedicated service.