Linuxcluster: Hardware

Hardware configuration

The HPC cluster at TUHH-RZ consists of 200 compute nodes, several login nodes and a parallel storage system with a capacity of 300TB. All in all, about 8700 CPU cores, 50 TByte RAM and some GPUs are available for compute intensive workloads.

Login nodes

The HPC cluster has several login nodes. Some login nodes may be temporarily unavailable due to maintenance. If you do not have specific hardware or software requirements you are advised to use the alias hpclogin.rz.tuhh.de.
Nodes Cores CPU Type RAM Recommended usage
hpc2.rz.tuhh.de 2× 16 2× AMD Epyc 9124 384 GB managing batch jobs, data transfer, building software,
pre- and postprocessing, short test runs
hpc3.rz.tuhh.de 2× 16 2× AMD Epyc 9124 384 GB managing batch jobs, data transfer, building software,
pre- and postprocessing, short test runs
hpc4.rz.tuhh.de 2× 10 2× Intel Xeon E5-2660v3 128 GB managing batch jobs, data transfer, building software,
pre- and postprocessing, short test runs
hpc5.rz.tuhh.de 2× 10 2× Intel Xeon E5-2660v3 128 GB managing batch jobs, data transfer, building software,
pre- and postprocessing, short test runs

Compute nodes

Nodes Cores CPU Type RAM Comment
n[001-056] 2× 32 2× AMD Epyc 9354 384 GB
n[057-112] 2× 32 2× AMD Epyc 9354 768 GB
g[209-216] 2× 14 2× Intel E5-2680v4 128 GB
g[217-224] 2× 16 2× Intel Xeon Gold 6130 192 GB
g[225-228] 2× 24 2× Intel Xeon Gold 5318Y 512 GB
u[008-009] 2× 36 2× Xeon Platinum 8352V 512 GB With four NVidia Tesla A100
(80GB Memory each)
u[010-011] 2× 32 2× AMD Epyc 9334 768 GB With four NVidia Tesla H100
(80GB Memory each)
 

Software

Storage

The HPC cluster provides several storage classes for different purposes.
HOME WORK SCRATCH PERM
accessible via environment variable $HOME $WORK $SCRATCH $PERM
mounted at /fibus /work /usertemp /perm
local / network file system network file system (NFS) network file system (Lustre) local file system network file system (NFS)
purpose Software, scripts, small files that need a good backup I/O-intensive data, pre- and post-processing I/O-intensive data during simulation data with medium lifetime and low I/O
Default quota 20 GB 5 TB / 10M inodes none, but limited by size of local medium 1 TB
Data lifetime until HPC access expires 2 months until reboot of server until HPC access expires
Characteristics Permissions may be set to share data with other users. On compute nodes just read-only access.
Snapshots yes, at ~/.snapshot/ no no yes, at /perm/.zfs/snapshot/