Ucd sonic

.github/workflows/main.yml

@@ -100,6 +100,7 @@ jobs:
           - "seawulf"
           - "seg_globe"
           - "tigem"
+          - "ucd_sonic"
           - "ucl_myriad"
           - "uct_hpc"
           - "uge"

README.md

@@ -155,6 +155,7 @@ Currently documentation is available for the following systems:
 - [SEG_GLOBE](docs/seg_globe.md)
 - [Super Computing Wales](docs/scw.md)
 - [TIGEM](docs/tigem.md)
+- [UCD_SONIC](docs/ucd_sonic.md)
 - [UCL_MYRIAD](docs/ucl_myriad.md)
 - [UCT_HPC](docs/uct_hpc.md)
 - [UGE](docs/uge.md)

conf/ucd_sonic.config

@@ -0,0 +1,38 @@
+//UCD_SONIC config
+params {
+  config_profile_name = 'UCD_SONIC'
+  config_profile_description = 'University College Dublin Sonic HPC profile provided by nf-core/configs.'
+  config_profile_contact = 'Bruce Moran (@brucemoran)'
+  config_profile_url = 'https://www.ucd.ie/itservices/ourservices/researchit/researchcomputing/sonichpc/'
+  //max_memory = 1.MB //yes really, allocation made based on other factors and cannot be done via params.max_memory...
+  max_cpus = 40
+  max_time = 12.h
+}
+
+process {
+  executor = 'slurm'
+  queue = 'shared'
+  queueSize = 50
+  submitRateLimit = '10 sec'
+  maxRetries = 2
+  beforeScript = 'export NXF_OPTS="-Xms2G -Xmx40G"; module load nextflow/22.04.5.5708 singularity/3.5.2'
+  clusterOptions = { "--mem 1M" } 
+  cache = 'lenient'
+  memory = 1.MB
+}
+
+cleanup = true
+
+profiles {
+  moranb {
+    singularity {
+      enabled = true
+      autoMounts = true
+      cacheDir = '/home/people/moranb/singularity/nextflow'
+      runOptions = '--bind /home/people/moranb,/scratch/moranb'
+    }
+    env {
+      NXF_OPTS="-Xms2G -Xmx40G"
+    }
+  }
+}

docs/ucd_sonic.md

@@ -0,0 +1,18 @@
+# nf-core/configs: UCD Sonic Configuration
+
+The nf-core pipeline [sarek](https://nf-co.re/sarek) has been successfully tested on [Sonic HPC] (https://www.ucd.ie/itservices/ourservices/researchit/researchcomputing/sonichpc) using this config.
+
+This Sonic config offers very minimal options to users but should allow the use of all nf-core pipelines.
+
+## sabtch_nxf_creator
+To run a pipeline, you should use the [sbatch_nxf_creator](https://github.com/brucemoran/sbatch_nxf_creator) method.
+
+This requires you to write a small YAML file including all of your sample names that you want to process individually, as well as your basic Nextflow command and where the output should be written.
+
+There is some weirdness on Sonic via SLURM (maybe it's just me!) which required me to write this method, maybe you can bypass it.
+
+The `sbatch_nxf_creator.sh` script takes you YAML and makes an `sbatch` file that can be submitted to the scheduler via a simple `sbatch my_file.sbatch` command.
+
+Practically, it means appending `srun` to the command and appending `-with-mpi`, as well as a few other vaguaries like specifying output and loading `Nextflow` and `Singularity`.  
+
+Even more practical is that as new samples arrive to you, you can add them to your YAML and then create new sbatch files, which I think is good.

nfcore_custom.config

@@ -82,6 +82,7 @@ profiles {
     seawulf          { includeConfig "${params.custom_config_base}/conf/seawulf.config"}
     seg_globe        { includeConfig "${params.custom_config_base}/conf/seg_globe.config"}
     tigem            { includeConfig "${params.custom_config_base}/conf/tigem.config"}
+    ucd_sonic        { includeConfig "${params.custom_config_base}/conf/ucd_sonic.config"}
     ucl_myriad       { includeConfig "${params.custom_config_base}/conf/ucl_myriad.config"}
     uct_hpc          { includeConfig "${params.custom_config_base}/conf/uct_hpc.config" }
     uge              { includeConfig "${params.custom_config_base}/conf/uge.config" }