issues
search
umccr
/
nextflow-stack
AWS stack to run Nextflow pipelines on Batch using shared resources
0
stars
0
forks
source link
Configure resource ramping for oncoanalyser
#39
Closed
scwatts
closed
1 year ago
scwatts
commented
1 year ago
Background
samples with many SVs can require far greater amounts of memory than typical samples
rather than inefficiently over allocating memory for most samples, resource ramping is now used
Changes
failed tasks are automatically retried with increased memory and reduced number of CPUs
this is currently limited to a maximum of one retry per task
a new queue is available specifically for memory ramping,
nextflow-task-16cpu_128gb-spot
ramping is applied only to oncoanalyser SvPrep, GRIDSS, and PURPLE processes
Background
Changes
nextflow-task-16cpu_128gb-spot