Athena cluster

Athena cluster

Cluster information and best practices

We recommend you watch this video explaining the cluster, how it works, and how users can use it optimally.

Hardware information

The athena cluster runs Rocky Linux 9 and comprises compute nodes organized in a 2-level hierarchy of partitions. The lustre storage filesystem includes 3 PB of shared space.

Consult detailed information about the hardware specifications of each node.

Consult the real-time dashboard with the current utilization of the cluster. Review the FAQ with the most common do's/don'ts, questions, and errors.

Partition L1

Partition L2

# Nodes

CPU model

# CPU cores per node

GPUs

RAM per node

Local /tmp per node

Partition L1

Partition L2

# Nodes

CPU model

# CPU cores per node

GPUs

RAM per node

Local /tmp per node

cpu

cpu-small

73

Intel Xeon

28 - 36

-

96 - 256 GB

2 TB SSD

cpu-large

43

AMD EPYC

128

-

256 - 1500 GB

2 -4 TB NVMe















gpu

gpu-v100

4

Intel / AMD

32 - 128

2x NVIDIA V100 16 / 32 GB

384 - 512 GB

2 TB SSD

gpu-a100

2

AMD EPYC

128

4x NVIDIA A100 SMX 80 GB

1024 GB

16 TB NVMe

gpu-h100

3

Intel Xeon

64 - 112

4-8x NVIDIA H100 SMX 80 GB

1024-2048 GB

14-25 TB NVMe

Total resources

9218 CPU cores

36 GPUs

66.8 TB RAM



2 CPUs and 8 GB of RAM are reserved on each node for the operating system and they cannot be allocated to users in Slurm.

Software information

All nodes provide basic software provided by default in Rocky 9, including GCC 11, Java 11 (openjdk), and Python 3.9

For Python, we recommend creating virtual environments for your projects. These enable you to keep the correct versions of Python packages you install using pip. Always use pip install --user to install the packages in your home directory.

Additional software and versions are provided via modules. Click on the software name for instructions. Use the command "module avail" to see all modules. Use the command "module load modulename" to load a specific module. Please read the documentation on how to use modules to load your software version.

Software

Versions

GPU-accelerated

Software

Versions

GPU-accelerated

AMS

2023.104

No

Anaconda3

2023.09-0

No

Ansys

2024R2

Some functions

Apptainer

1.2.4

No

AWS CLI

2.15.48

No

bcftools

1.18, 1.21

No

Bedtools

2.31.0

No

Bioperl

1.7.8

No

Boost

1.83.0

No

Bowtie2

2.5.2

No

BrainGlobe

1.3.1

Yes

BWA

0.7.17-r1198

No

CFDEM

3.8

No

Cell Ranger

6.0.2, 7.2.0, 8.0.1, 9.0.1, 10.1.0

No

Cell Ranger ARC

2.0.2

No

Cell Ranger ATAC

2.1.0

No

Chimera

1.17.3

Yes

ChimeraX

1.7.1

Yes

Cmake

3.27.7

No

Coot

1.1.09

Yes

CUDA

12.3

Yes

Cutadapt

4.9

No

CryoSPARC

4.5.1

Yes

CTFFIND

4.1.14

No

Dalton

2020.1

No

demuxlet

2021.02

No

DeepLabCut

3.0.0

Yes

Elmer

9.0

No

Easybuild

4.8.2

No

fastp

0.23.4

No

FastQC

0.12.1

No

Gaussian16

C.01

No

GCC

9.5.0, 10.5.0

No

GCC Toolset

12, 13

No

GffCompare

0.12.6

No

Go

1.21.4

No

GOLD

2022.1

No

Google Cloud CLI

2024.10

No

GROMACS

2023.3

Yes

GSL

2.7.1

No

Gurobi

11.0.0

No

HDF5

1.14.3

No

HISAT

2.2.1

No

htslib

1.18, 1.21

No

Intel oneAPI

2024.0.0

No

IQ-TREE

2.2.2.6

No

JAGS

4.3.2

No

Jellyfish

2.3.1

No

Julia

1.9.2

No

Matlab

R2024a

Some functions

MCNP

6.2

No

MetaPhlAn

4.1.1

No

Miniconda3

23.9.0

No

Miniforge3

23.3.1

No

MotionCor2

1.6.4

Yes

MotionCor3

1.1.1

Yes

MultiQC

1.17

No

Muscle

5.1.0

No

NAMD

2.14, 3.0b5

Yes

ncbi-blast

2.15.0

No

nekRS

23.0

Yes

Nextflow

23.10.0

No

NWChem

7.2.2

No

Open Babel

3.1.1

No

OpenCV

4.8.1

No

OpenFOAM

5, 8, 9, 10, 11

No

OpenJDK

17.0.2, 21.0.1

No

OpenMC

0.14.0

No

OpenMPI

4.1.6

No

ORCA

5.0.4, 6.0.0

No

OrthoFinder

2.5.4

No

PAML

4.10.7

No

PGI

2018

No

Phenix

1.21.1-5286

No

Picard

3.1.0