Restarting from a checkpoint should restore the time step?

[qingli411]

It looks like that restarting from a checkpoint is not bit-for-bit? I think the issue is that when restarting from a checkpoint the time step is not restored -- it is still using the initial time step defined in Simulation(), not the last_Δt from the saved Clock object in the checkpoint file. See the example below, which is the output of the attached test case.

In the pickup run I changed the onscreen output from every 10 iterations to every 1 iterations to see the time step. Rather than using the previous time step (5.973 s) from the checkpoint, the pickup run is using a time step of 10 s which is the value when defining simulation = Simulation(model, Δt=10, stop_iteration=220). I’m using v0.91.5.

Initial run

[ Info: Initializing simulation...
Iteration: 0000, time: 0 seconds, Δt: 11 seconds, max(|u|) = 2.5e-01 ms⁻¹, wall time: 0 seconds
[ Info:     ... simulation initialization complete (13.909 seconds)
[ Info: Executing initial time step...
[ Info:     ... initial time step complete (4.696 seconds).
Iteration: 0010, time: 1.833 minutes, Δt: 11.212 seconds, max(|u|) = 2.6e-01 ms⁻¹, wall time: 19.116 seconds
Iteration: 0020, time: 3.702 minutes, Δt: 10.681 seconds, max(|u|) = 2.8e-01 ms⁻¹, wall time: 19.345 seconds
Iteration: 0030, time: 5.482 minutes, Δt: 10.215 seconds, max(|u|) = 2.9e-01 ms⁻¹, wall time: 19.629 seconds
Iteration: 0040, time: 7.185 minutes, Δt: 9.802 seconds, max(|u|) = 3.0e-01 ms⁻¹, wall time: 19.854 seconds
Iteration: 0050, time: 8.819 minutes, Δt: 9.433 seconds, max(|u|) = 3.1e-01 ms⁻¹, wall time: 20.082 seconds
Iteration: 0060, time: 10.391 minutes, Δt: 9.100 seconds, max(|u|) = 3.2e-01 ms⁻¹, wall time: 20.306 seconds
Iteration: 0070, time: 11.907 minutes, Δt: 8.798 seconds, max(|u|) = 3.3e-01 ms⁻¹, wall time: 20.559 seconds
Iteration: 0080, time: 13.374 minutes, Δt: 8.523 seconds, max(|u|) = 3.4e-01 ms⁻¹, wall time: 20.773 seconds
Iteration: 0090, time: 14.794 minutes, Δt: 8.270 seconds, max(|u|) = 3.5e-01 ms⁻¹, wall time: 21.001 seconds
Iteration: 0100, time: 16.173 minutes, Δt: 8.036 seconds, max(|u|) = 3.6e-01 ms⁻¹, wall time: 21.230 seconds
Iteration: 0110, time: 17.512 minutes, Δt: 7.820 seconds, max(|u|) = 3.7e-01 ms⁻¹, wall time: 21.456 seconds
Iteration: 0120, time: 18.815 minutes, Δt: 7.617 seconds, max(|u|) = 3.8e-01 ms⁻¹, wall time: 21.688 seconds
Iteration: 0130, time: 20.085 minutes, Δt: 7.424 seconds, max(|u|) = 3.9e-01 ms⁻¹, wall time: 21.914 seconds
Iteration: 0140, time: 21.322 minutes, Δt: 7.238 seconds, max(|u|) = 4.0e-01 ms⁻¹, wall time: 22.140 seconds
Iteration: 0150, time: 22.528 minutes, Δt: 7.046 seconds, max(|u|) = 4.0e-01 ms⁻¹, wall time: 22.366 seconds
Iteration: 0160, time: 23.703 minutes, Δt: 6.792 seconds, max(|u|) = 4.1e-01 ms⁻¹, wall time: 22.597 seconds
Iteration: 0170, time: 24.835 minutes, Δt: 6.500 seconds, max(|u|) = 4.2e-01 ms⁻¹, wall time: 22.822 seconds
Iteration: 0180, time: 25.918 minutes, Δt: 6.230 seconds, max(|u|) = 4.3e-01 ms⁻¹, wall time: 23.047 seconds
Iteration: 0190, time: 26.956 minutes, Δt: 5.973 seconds, max(|u|) = 4.3e-01 ms⁻¹, wall time: 23.271 seconds
Iteration: 0200, time: 27.952 minutes, Δt: 5.930 seconds, max(|u|) = 4.3e-01 ms⁻¹, wall time: 23.504 seconds
Iteration: 0210, time: 28.940 minutes, Δt: 5.979 seconds, max(|u|) = 4.1e-01 ms⁻¹, wall time: 24.139 seconds
[ Info: Simulation is stopping after running for 24.362 seconds.
[ Info: Model iteration 220 equals or exceeds stop iteration 220.
Iteration: 0220, time: 29.937 minutes, Δt: 5.898 seconds, max(|u|) = 4.1e-01 ms⁻¹, wall time: 24.362 seconds

Pickup run

[ Info: Initializing simulation...
[ Info:     ... simulation initialization complete (51.655 ms)
[ Info: Executing initial time step...
[ Info:     ... initial time step complete (5.017 seconds).
Iteration: 0201, time: 28.118 minutes, Δt: 10 seconds, max(|u|) = 4.2e-01 ms⁻¹, wall time: 0 seconds
Iteration: 0202, time: 28.285 minutes, Δt: 10 seconds, max(|u|) = 4.2e-01 ms⁻¹, wall time: 10.477 seconds
Iteration: 0203, time: 28.452 minutes, Δt: 10 seconds, max(|u|) = 4.1e-01 ms⁻¹, wall time: 10.499 seconds
Iteration: 0204, time: 28.618 minutes, Δt: 10 seconds, max(|u|) = 4.1e-01 ms⁻¹, wall time: 10.520 seconds
Iteration: 0205, time: 28.785 minutes, Δt: 10 seconds, max(|u|) = 4.1e-01 ms⁻¹, wall time: 10.544 seconds
Iteration: 0206, time: 28.952 minutes, Δt: 10 seconds, max(|u|) = 4.1e-01 ms⁻¹, wall time: 10.570 seconds
Iteration: 0207, time: 29.118 minutes, Δt: 10 seconds, max(|u|) = 4.1e-01 ms⁻¹, wall time: 10.595 seconds
Iteration: 0208, time: 29.285 minutes, Δt: 10 seconds, max(|u|) = 4.1e-01 ms⁻¹, wall time: 10.621 seconds
Iteration: 0209, time: 29.452 minutes, Δt: 10 seconds, max(|u|) = 4.1e-01 ms⁻¹, wall time: 10.647 seconds
Iteration: 0210, time: 29.618 minutes, Δt: 6.183 seconds, max(|u|) = 4.0e-01 ms⁻¹, wall time: 10.673 seconds
Iteration: 0211, time: 29.722 minutes, Δt: 6.183 seconds, max(|u|) = 4.0e-01 ms⁻¹, wall time: 14.392 seconds
Iteration: 0212, time: 29.825 minutes, Δt: 6.183 seconds, max(|u|) = 4.0e-01 ms⁻¹, wall time: 14.413 seconds
Iteration: 0213, time: 29.928 minutes, Δt: 6.183 seconds, max(|u|) = 4.0e-01 ms⁻¹, wall time: 14.435 seconds
Iteration: 0214, time: 30.031 minutes, Δt: 6.183 seconds, max(|u|) = 4.0e-01 ms⁻¹, wall time: 14.459 seconds
Iteration: 0215, time: 30.134 minutes, Δt: 6.183 seconds, max(|u|) = 4.0e-01 ms⁻¹, wall time: 14.485 seconds
Iteration: 0216, time: 30.237 minutes, Δt: 6.183 seconds, max(|u|) = 4.0e-01 ms⁻¹, wall time: 14.510 seconds
Iteration: 0217, time: 30.340 minutes, Δt: 6.183 seconds, max(|u|) = 4.0e-01 ms⁻¹, wall time: 14.536 seconds
Iteration: 0218, time: 30.443 minutes, Δt: 6.183 seconds, max(|u|) = 4.0e-01 ms⁻¹, wall time: 14.559 seconds
Iteration: 0219, time: 30.546 minutes, Δt: 6.183 seconds, max(|u|) = 4.0e-01 ms⁻¹, wall time: 14.582 seconds
[ Info: Simulation is stopping after running for 14.606 seconds.
[ Info: Model iteration 220 equals or exceeds stop iteration 220.
Iteration: 0220, time: 30.649 minutes, Δt: 6.023 seconds, max(|u|) = 4.0e-01 ms⁻¹, wall time: 14.606 seconds

run.jl.zip

@glwagner