3.1.10. RunControl

Option aliases:

  • CAPEFile ? JSONFile

  • ConcurrentPythonFuncs ? WorkerPythonFuncs

  • ConcurrentShellCmds ? WorkerShellCmds

  • Continue ? ContinuePhase

  • GPU ? gpu

  • NSubProc ? Maxforks

  • NSubProcess ? MaxForks

  • PostCmds ? PostShellCmds

  • PostFuncs ? PostPythonFuncs

  • PostPyFuncs ? PostPythonFuncs

  • PreCmds ? PreShellCmds

  • PreFuncs ? PrePythonFuncs

  • PrePyFuncs ? PrePythonFuncs

  • Resubmit ? ResubmitNextPhase

  • WatcherCmds ? WorkerShellCmds

  • WatcherFuncs ? WorkerPythonFuncs

  • WatherTimeout ? WorkerTimeout

  • WorkerCmds ? WorkerShellCmds

  • WorkerFuncs ? WorkerPythonFuncs

  • WorkerMaxWaitTime ? WorkerTimeout

  • WorkerTimeOut ? WorkerTimeout

  • nJob ? NJob

  • sbatch ? slurm

Recognized options:

BatchSize: {1000} | int

number of iters to store in single batch

ContinuePhase: {None} | True | False

whether restarts of same phase can use same job

JSONFile: {None} | str

name of JSON file from which settings originated

MPI: {False} | True

whether or not to run MPI in phase

MaxForks: {4} | int

maximum forks to use for parallel post-processing

MaxWorkers: {8} | int

maximum threads to use for parallel post-processing

NJob: {0} | int

number of jobs to run concurrently

PhaseIters: {None} | int

check-point iterations for phase j

PhaseSequence: {None} | int

list of phase indices to run

PhaseWatcherCmds: {None} | object

value of option “PhaseWatcherCmds”

PhaseWatcherFuncs: {None} | object

value of option “PhaseWatcherFuncs”

PostPythonFuncs: {None} | list[dict | str]

Python functions to run after each cycle

PostShellCmds: {None} | list[dict | str]

list of commands to run after each cycle

PreMesh: {False} | True

option to generate mesh before submitting job

PrePythonFuncs: {None} | list[dict | str]

Python functions to run before each cycle

PreShellCmds: {None} | list[dict | str]

list of commands to run before each cycle

RestartSamePhase: {True} | False

whether to restart same phase if needed

ResubmitNextPhase: {False} | True

whether new job submits at end of phase

ResubmitSamePhase: {False} | True

whether same-phase repeats need new job

RootDir: {None} | str

(absolute) base folder from which CAPE settings were read

StartNextPhase: {True} | False

whether to start next phase or stop CAPE

Verbose: {False} | True

“RunControl” verbosity flag

WarmStart: {False} | True

whether to warm start a case

WarmStartFolder: {None} | str

folder from which to get warm-start file

WarmStartPhase: {None} | int

phase from which to warm-start a case

WorkerPythonFuncs: {None} | list[dict | str]

functions to run concurrently during phase

WorkerShellCmds: {None} | list[str]

shell commands to run concurrently during phase

WorkerSleepTime: {10.0} | int | float

value of option “WorkerSleepTime”

WorkerTimeout: {600.0} | int | float

max time to wait for workers after CFD phase

ZombieFiles: {None} | list[str]

file name flobs to check mod time for zombie status

ZombieTimeout: {45.0} | int | float

minutes to wait before considering a case a zombie

gpu: {None} | True | False

Whether to use GPU-version of code, if separate

mpicmd: 'mpiexec' | 'mpirun'

MPI executable name

nIter: {None} | int

number of iterations to run in phase j

nProc: {None} | int | float

number (or fraction) of cores to use (or omit if negative)

qsub: {False} | True

whether or not to submit jobs with PBS

quickstart: {False} | True

option to generate mesh before submitting job

slurm: {False} | True

whether or not to submit jobs with Slurm

Subsections: