nf-core/configs: NC State Hazel HPC Configuration
nf-core pipelines have been configured for use on the Hazel HPC at NC State University.
To use, run the pipeline with -profile hazel.
Example:
nextflow run nf-core/<pipeline> -profile hazel
Running this command will download the hazel config from the nf-core repository, and will submit nextflow processes as jobs to the slurm job scheduler. Therefore, the nextflow run command must be run from the login node. Using this profile, a docker image containing all of the required software will be downloaded, and converted to a Singularity image before execution of the pipeline. These images are cached in /share/$GROUP/$USER/tmp so they don’t have to be re-downloaded on subsequent runs.
Loading required environment modules
Before running the pipeline you will need to load Nextflow using the environment module system on hazel. Some of these modules are managed by the BRC, and are not in the default module path. See this resource for information about using these modules: https://hurwitzlab.github.io/COS_Compute_Handbook/chapters/14_loading_brc_modules.html
## Load Nextflow environment modules
module purge # Optional, unload any currently loaded modules.
module load nextflow/26.04.3 # Required
module load nf-core # Optional, manages and downloads nf-core pipelines for common use. Requires singularity.Job submission partitions
By default, using this profile will submit jobs to the compute partition. To specify a different partition for submission, use --partition <PARTITION NAME> in the nextflow run command.
Nextflow processes with the process_gpu label will be submitted to the gpu partition by default; override with --gpu_partition <PARTITION NAME>. To request a different GPU type or multiple GPUs, use --hazel_gpu 'gpu:<gpu-type>:<num-gpus>' at runtime.
Overriding Hazel config options
Nextflow-level options passed to your nextflow run command will override options contained in the Hazel institutional profile:
nextflow run nf-core/<pipeline> -profile hazel --max_memory 256.GB # raise memory ceiling
nextflow run nf-core/<pipeline> -profile hazel --hazel_gpu 'gpu:a100:2' # different GPUFor persistent overrides, use a local config file with -c my_overrides.config.
You will need an account to use the Hazel HPC in order to run the pipeline. If in doubt contact IT or your PI to get access.
Nextflow will need to submit the jobs via the job scheduler to the HPC cluster and as such the commands above will have to be executed on one of the login nodes. If in doubt contact IT.
Config file
// NC State University - Hazel HPC Cluster
// nf-core institutional config
// Scheduler: SLURM | Containers: Singularity/Apptainer
//
// To activate: nextflow run <pipeline> -profile hazel
//
// Submit an issue or PR to add this to the nf-core/configs repo:
// https://github.com/nf-core/configs
params {
config_profile_name = 'NC State University - Hazel HPC'
config_profile_description = 'nf-core config for the Hazel HPC cluster at NC State University'
config_profile_contact = 'Seth Weaver (@sweavs111)'
config_profile_url = 'https://brc.ncsu.edu/'
max_memory = 128.GB
max_cpus = 24
max_time = 120.h
// Default GPU allocation. Override at runtime with --hazel_gpu 'gpu:<type>:<num>'
hazel_gpu = 'gpu:h100:1'
// Default partitions. Override at runtime with --partition or --gpu_partition
partition = 'compute'
gpu_partition = 'gpu'
}
validation {
ignoreParams = ['hazel_gpu', 'partition', 'gpu_partition']
}
// ─── Singularity ────────────────────────────────────────────────────────────
singularity {
enabled = true
autoMounts = true
cacheDir = System.getenv('APPTAINER_CACHEDIR') ?: "/share/${System.getenv('GROUP')}/${System.getenv('USER')}/tmp"
}
// ─── Executor tuning ────────────────────────────────────────────────────────
executor {
queueSize = 16
pollInterval = '1 min'
queueStatInterval = '5 min'
submitRateLimit = '10 sec'
}
// ─── Process defaults ────────────────────────────────────────────────────────
process {
beforeScript = 'module load singularity'
executor = 'slurm'
queue = params.partition
resourceLimits = [
memory: params.max_memory,
cpus: params.max_cpus,
time: params.max_time
]
// GPU jobs use gpu partition
withLabel: 'process_gpu' {
queue = params.gpu_partition
clusterOptions = { "--gres=${params.hazel_gpu}" }
}
}