Restarting a steady state computation does not reproduce same residuals.

[clarkpede]

Steps to reproduce

1. Run a steady state computation for 10 iterations.
2. Run a steady state computation for 10 iterations, split into two:

• Run a steady state computation for 5 iterations.
• Restart the steady state computation for 5 more iterations.

3. Compare the residuals for Steps 1 and 2. Slight differences (~1%) appear. The final fields are also slightly different.

This bug seems to originate from SU2's master branch (not our fork).

I'm attaching a file that goes over both the bug, and has scripts to reproduce the issue.
restart_steady_state.tar.gz