Thanks to the generous sponsorship from Nebius, we have access to 16 powerful H100 GPUs for the duration of this hackathon. Nebius is providing these resources through their Slurm-based workload manager called Soperator, enabling participants to test and develop GPU-accelerated tools with Nextflow using a familiar Slurm interface. We welcome people from all other projects to use these resources if they need access to GPUs.
To enable participants access to the Slurm cluster with the GPU resources, we need to add your public key to the authorized_keys list on the Slurm cluster. If you want access, please contact Florian with your public key.
You can create a dedicated ssh keypair with ssh-keygen like this, replacing the example email address with your own (store in a safe location):
ssh-keygen -t rsa -C "your_email@example.com"Once your public key has been added to the authorized_keys list on the cluster, you can connect to it with:
ssh -i /path/to/your/private_key username@hostnameImportant information: Access to the cluster and all data on the cluster will be wiped once the hackathon ends, make sure to transfer all important data from the Slurm cluster to your own storage before the hackathon ends!
Goal
Use GPUs via Slurm to test and develop GPU accelerated tools in Nextflow and nf-core.
