PerfGraph output misses the time sink in NS solves

[GiudGiud]

Bug Description

See this log

      Finished Computing Residual                                                        [  7.39 s] [  -62 MB]
    Finished Computing Initial Residual                                                  [  7.39 s] [  -62 MB]
    Finished Computing Residual                                                          [  5.66 s] [  -40 MB]
 0 Nonlinear |R| = 1.508806e+06
    Finished Computing Jacobian                                                          [  7.95 s] [  218 MB]
    Finished Computing Residual                                                          [  5.68 s] [    9 MB]
 1 Nonlinear |R| = 3.109142e+03
    Computing Jacobian..................                                                 [ 99.21 s] [ -396 MB]
    Finished Computing Residual                                                          [  5.73 s] [  -27 MB]
 2 Nonlinear |R| = 9.077032e+00
    Finished Computing Jacobian                                                          [  7.39 s] [ -102 MB]
    Finished Computing Residual                                                          [  5.90 s] [   -8 MB]
 3 Nonlinear |R| = 1.494024e-03
 Solve Converged!
  Finished Solving                                                                       [609.70 s] [ -613 MB]

we are not printing dot dot dots at the right time, and we dont show most of the time spent. So some section must be uncovered

Steps to Reproduce

run neams-th/sfr/abtr/model2 inputs with 8 processes

Impact

This isnt really affecting me. But we would have an easier time understanding the time sinks if perf graph console singled them out. I m going to post further details with perfgraphoutput to get the full picture hopefully

[GiudGiud]

I'm thinking the RC loop is not captured by the perf graph. I m sure if Alex or Peter see this ticket they will immediately know.

[GiudGiud]

yeah possibly we just need to add a perf graph section to INSFVRhieChowInterpolator::execute()