It should be very helpful to allow users to set them properly according to what the worker VM really has.
The idea is to let user set max number of CPU cores and memory a worker compute node has. Such information will be used by internal individual steps to decide how many actual processes/threads will be used taking into account both available CPU core and memory.
junjun-zhang
commented
5 years ago
Currently these two key parameters are hardcoded: https://github.com/ICGC-TCGA-PanCancer/CGP-Somatic-Docker/blob/develop/scripts/run_seqware_workflow.py#L108-L109.
It should be very helpful to allow users to set them properly according to what the worker VM really has.
The idea is to let user set max number of CPU cores and memory a worker compute node has. Such information will be used by internal individual steps to decide how many actual processes/threads will be used taking into account both available CPU core and memory.