Changes¶
The format of the checkpoint files has been updated.
Before, the restart parameters write written to checkpoint.chk.par. The old behaviour of ommitting the “.par” for the checkpoint has been reintroduced. In addition, all of the state variables are now saved in Python Dill (Pickle) format in a “.chk.pkl” file. To restart, use the “.chk” file as input, for example:
srun pkdgrav3 checkpoint.00010.chk
To override parameters you can add them as arguments to the restore call.
For example if you wanted to turn on parallel reading and update
the number of readers to 100 you would change:
msr.restore(__file__)
to:
msr.restore(__file__,bParaRead=True,nParaRead=100)
The new values of those parameters will persist on subsequent restarts.
The parameter processing has undergone significant changes.
- dTheta, dTheta2, dTheta20
Only dTheta remains. To retain the old behaviour see the accuracy extension.
- dxPeriod, dyPeriod, dzPeriod
These parameters have been removed. Instead, use dPeriod and set it to a list of three values.
- hxLCP,hyLCP,hzLCP
These parameters have been replaced with hLCP, a vector of three values.
- achOutTimes
This parameter has been removed. The functionality can be duplicated by using the list format for nSteps and dRedTo. For example, if you have a list of redshift to output, you can set the number of steps to be one for each interval:
dRedFrom = 49 # Start at z=49 dRedTo = [10,2,1,0.5,0] # Step to z=10, 2, 1, 0.5 and 0 nSteps = [1] * len(dRedTo) # Taking one step for each interval iOutInterval = 1 # Output every interval
If you have expansion factors you can just convert them to redshift:
dRedTo = [1 / a - 1 for a in [0.1,0.5,1] ]
- nStepsSync,dRedSync
These parameter have been removed. The same functionality is now more generally available with nSteps and dRedTo. For example, before you might have had:
dRedFrom = 49 dRedTo = 0 nSteps = 160 nStepsSync = 60 dRedSync = 10
This would take 60 steps to redshift 10, then 100 steps to redshift 0. The new way of specifying this is:
dRedFrom = 40 dRedTo = [10,0] nSteps = [60,100]