JASMIN Help Site logo JASMIN Help Site logo
  • Docs 
  • Guides 
  • Training 
  • Discussions   

  •   Search this site  

Can't find what you're looking for?

Try our Google custom search, across all JASMIN sites

Docs
  • getting started
    • get started with jasmin
    • generate ssh key pair
    • get jasmin portal account
    • get login account
    • beginners training workshop
    • how to contact us about jasmin issues
    • jasmin status
    • jasmin training accounts
    • tips for new users
    • how to login
    • multiple account types
    • present ssh key
    • reconfirm email address
    • reset jasmin account password
    • ssh auth
    • storage
    • understanding new jasmin storage
    • update a jasmin account
  • interactive computing
    • interactive computing overview
    • check network details
    • login servers
    • login problems
    • graphical linux desktop access using nx
    • sci servers
    • tenancy sci analysis vms
    • transfer servers
    • jasmin notebooks service
    • jasmin notebooks service with gpus
    • creating a virtual environment in the notebooks service
    • project specific servers
    • dask gateway
    • access from vscode
  • batch computing
    • lotus overview
    • slurm scheduler overview
    • slurm queues
    • lotus cluster specification
    • how to monitor slurm jobs
    • how to submit a job
    • how to submit an mpi parallel job
    • example job 2 calc md5s
    • orchid gpu cluster
    • slurm status
    • slurm quick reference
  • software on jasmin
    • software overview
    • quickstart software envs
    • python virtual environments
    • additional software
    • community software esmvaltool
    • community software checksit
    • compiling and linking
    • conda environments and python virtual environments
    • conda removal
    • creating and using miniforge environments
    • idl
    • jasmin sci software environment
    • jasmin software faqs
    • jaspy envs
    • matplotlib
    • nag library
    • name dispersion model
    • geocat replaces ncl
    • postgres databases on request
    • running python on jasmin
    • running r on jasmin
    • rocky9 migration 2024
    • share software envs
  • data transfer
    • data transfer overview
    • data transfer tools
    • globus transfers with jasmin
    • bbcp
    • ftp and lftp
    • globus command line interface
    • globus connect personal
    • gridftp ssh auth
    • rclone
    • rsync scp sftp
    • scheduling automating transfers
    • transfers from archer2
  • short term project storage
    • apply for access to a gws
    • elastic tape command line interface hints
    • faqs storage
    • gws etiquette
    • gws scanner ui
    • gws scanner
    • gws alert system
    • install xfc client
    • xfc
    • introduction to group workspaces
    • jdma
    • managing a gws
    • secondary copy using elastic tape
    • share gws data on jasmin
    • share gws data via http
    • using the jasmin object store
    • configuring cors for object storage
  • long term archive storage
    • ceda archive
  • mass
    • external access to mass faq
    • how to apply for mass access
    • moose the mass client user guide
    • setting up your jasmin account for access to mass
  • for cloud tenants
    • introduction to the jasmin cloud
    • jasmin cloud portal
    • cluster as a service
    • cluster as a service kubernetes
    • cluster as a service identity manager
    • cluster as a service slurm
    • cluster as a service pangeo
    • cluster as a service shared storage
    • adding and removing ssh keys from an external cloud vm
    • provisioning tenancy sci vm managed cloud
    • sysadmin guidance external cloud
    • best practice
  • workflow management
    • rose cylc on jasmin
    • using cron
  • uncategorized
    • mobaxterm
    • requesting resources
    • processing requests for resources
    • acknowledging jasmin
    • approving requests for access
    • working with many linux groups
    • jasmin conditions of use
  • getting started
    • get started with jasmin
    • generate ssh key pair
    • get jasmin portal account
    • get login account
    • beginners training workshop
    • how to contact us about jasmin issues
    • jasmin status
    • jasmin training accounts
    • tips for new users
    • how to login
    • multiple account types
    • present ssh key
    • reconfirm email address
    • reset jasmin account password
    • ssh auth
    • storage
    • understanding new jasmin storage
    • update a jasmin account
  • interactive computing
    • interactive computing overview
    • check network details
    • login servers
    • login problems
    • graphical linux desktop access using nx
    • sci servers
    • tenancy sci analysis vms
    • transfer servers
    • jasmin notebooks service
    • jasmin notebooks service with gpus
    • creating a virtual environment in the notebooks service
    • project specific servers
    • dask gateway
    • access from vscode
  • batch computing
    • lotus overview
    • slurm scheduler overview
    • slurm queues
    • lotus cluster specification
    • how to monitor slurm jobs
    • how to submit a job
    • how to submit an mpi parallel job
    • example job 2 calc md5s
    • orchid gpu cluster
    • slurm status
    • slurm quick reference
  • software on jasmin
    • software overview
    • quickstart software envs
    • python virtual environments
    • additional software
    • community software esmvaltool
    • community software checksit
    • compiling and linking
    • conda environments and python virtual environments
    • conda removal
    • creating and using miniforge environments
    • idl
    • jasmin sci software environment
    • jasmin software faqs
    • jaspy envs
    • matplotlib
    • nag library
    • name dispersion model
    • geocat replaces ncl
    • postgres databases on request
    • running python on jasmin
    • running r on jasmin
    • rocky9 migration 2024
    • share software envs
  • data transfer
    • data transfer overview
    • data transfer tools
    • globus transfers with jasmin
    • bbcp
    • ftp and lftp
    • globus command line interface
    • globus connect personal
    • gridftp ssh auth
    • rclone
    • rsync scp sftp
    • scheduling automating transfers
    • transfers from archer2
  • short term project storage
    • apply for access to a gws
    • elastic tape command line interface hints
    • faqs storage
    • gws etiquette
    • gws scanner ui
    • gws scanner
    • gws alert system
    • install xfc client
    • xfc
    • introduction to group workspaces
    • jdma
    • managing a gws
    • secondary copy using elastic tape
    • share gws data on jasmin
    • share gws data via http
    • using the jasmin object store
    • configuring cors for object storage
  • long term archive storage
    • ceda archive
  • mass
    • external access to mass faq
    • how to apply for mass access
    • moose the mass client user guide
    • setting up your jasmin account for access to mass
  • for cloud tenants
    • introduction to the jasmin cloud
    • jasmin cloud portal
    • cluster as a service
    • cluster as a service kubernetes
    • cluster as a service identity manager
    • cluster as a service slurm
    • cluster as a service pangeo
    • cluster as a service shared storage
    • adding and removing ssh keys from an external cloud vm
    • provisioning tenancy sci vm managed cloud
    • sysadmin guidance external cloud
    • best practice
  • workflow management
    • rose cylc on jasmin
    • using cron
  • uncategorized
    • mobaxterm
    • requesting resources
    • processing requests for resources
    • acknowledging jasmin
    • approving requests for access
    • working with many linux groups
    • jasmin conditions of use
  1.   Batch Computing
  1. Home
  2. Docs
  3. Batch Computing
  4. Orchid GPU cluster

Orchid GPU cluster

 

Share via
JASMIN Help Site
Link copied to clipboard

Details of JASMIN's GPU cluster, ORCHID

On this page
GPU cluster spec   Request access to ORCHID   Test a GPU job   GPU interactive node   Software Installed on the GPU cluster (to be updated)  
 
Not yet reviewed for compatibility with the new cluster, April 2025. Please adapt using new submission instructions. This alert will be removed once updated.

GPU cluster spec  

The JASMIN GPU cluster is composed of 16 GPU nodes:

  • 14 x standard GPU nodes with 4 GPU Nvidia A100 GPU cards each
  • 2 x large GPU nodes with 8 Nvidia A100 GPU cards
ORCHID GPU cluster
ORCHID GPU cluster

Request access to ORCHID  

Before using ORCHID on JASMIN, you will need:

  1. An existing JASMIN account and valid jasmin-login access role: Apply here     
  2. Subsequently (once jasmin-login has been approved and completed), the orchid access role: Apply here     

The jasmin-login access role ensures that your account is set up with access to the LOTUS batch processing cluster, while the orchid role grants access to the special LOTUS partition used by ORCHID.

Holding the orchid role also gives access to the GPU interactive node.

Note: In the supporting info on the orchid request form, please provide details on the software and the workflow that you will use/run on ORCHID.

Test a GPU job  

Testing a job on the JASMIN ORCHID GPU cluster can be carried out in an interactive mode by launching a pseudo-shell terminal Slurm job from a JASMIN scientific server e.g. sci-vm-01:

srun --gres=gpu:1 --partition=orchid --account=orchid --qos=orchid --pty /bin/bash
srun: job 24096593 queued and waiting for resources
srun: job 24096593 has been allocated resources

At this point, your shell prompt will change to the GPU node gpuhost016, which is allocated for this interactive session on ORCHID. You will have the one GPU allocated at this shell, as requested:

nvidia-smi
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 550.90.07              Driver Version: 550.90.07      CUDA Version: 12.4     |
|-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA A100-SXM4-40GB          On  |   00000000:01:00.0 Off |                    0 |
|                   ...                   |           ...          |         ...          |

Note that for batch mode, a GPU job is submitted using the Slurm command sbatch:

sbatch --gres=gpu:1 --partition=orchid --account=orchid --qos=orchid gpujobscript.sbatch

or by adding the following preamble in the job script file

#SBATCH --partition=orchid
#SBATCH --account=orchid
#SBATCH --gres=gpu:1

Note 1: gpuhost015 and gpuhost016 are the two largest nodes with 64 CPUs and 8 GPUs.

Note 2: CUDA Version: 11.6

Note 3: The Slurm batch partition/queue orchid has a maximum runtime of 24 hours and the default runtime is 1 hour. The maximum number of CPU cores per user is limited to 8 cores. If the limit is exceeded then the job is expected to be in a pending state with the reason being QOSGrpCpuLimit

GPU interactive node  

There is an interactive GPU node gpuhost001.jc.rl.ac.uk, with the same spec as other Orchid nodes, that you can access via a login server to prototype and test your GPU code prior to running as a batch job:

ssh -A gpuhost001.jc.rl.ac.uk

Software Installed on the GPU cluster (to be updated)  

  • CUDA drivers 10.1, and CUDA libraries 10.0, 10.1 and 11.4
  • CUDA DNN (Deep Neural Network Library)
  • NVIDIA container runtime (see notes below)
  • NGC client (GPU software hub for NVIDIA)
  • Singularity 3.7.0 - which supports NVIDIA/GPU containers
  • SCL Python 3.6
Last updated on 2025-04-09 as part of:  Initial review of Orchid cluster (735afdfb7)
On this page:
GPU cluster spec   Request access to ORCHID   Test a GPU job   GPU interactive node   Software Installed on the GPU cluster (to be updated)  
Follow us

Social media & development

   

Useful links

  • CEDA Archive 
  • CEDA Catalogue 
  • JASMIN 
  • JASMIN Accounts Portal 
  • JASMIN Projects Portal 
  • JASMIN Cloud Portal 
  • JASMIN Notebooks Service 
  • JASMIN Community Discussions 

Contact us

  • Helpdesk
UKRI/STFC logo
UKRI/NERC logo
NCAS logo
NCEO logo
Accessibility | Terms and Conditions | Privacy and Cookies
Copyright © 2025 Science and Technology Facilities Council.
Hinode theme for Hugo licensed under Creative Commons (CC BY-NC-SA 4.0).
JASMIN Help Site
Code copied to clipboard