An Error related to "BoxToCell"

[friedenhe]

Discussed in #825

^{Originally posted by WangWen-kang May 7, 2025}
Dear Prof. He
I run into a problem when I test the Periodic_hill tutorial. I changed the sources of UFieldVar and BetaVar to "boxToCell". The related code excerpt is as following

    "function": {
        "UFieldVar": {
            "type": "variance",
            "source": "boxToCell",#Options are: allCells, patchToFace, or boxToCell
            "min": [0.0, 0.0, 0.0],
            "max": [3.0, 1.0, 1.0],
            "scale": 1.0,
            "mode": "field",
            "varName": "U",
            #"data":"Udata", the class search for varName+"data" Automatically
            "varType": "vector",
            "components": [0, 1],
            "timeDependentRefData": False,
        },
        "betaVar": {
            "type": "variance",
            "source": "boxToCell",#Options are: allCells, patchToFace, or boxToCell
            "min": [0.0, 0.0, 0.0],
            "max": [3.0, 1.0, 1.0],
            "scale": 1.0,
            "mode": "field",
            "varName": "betaFINuTilda",
            "varType": "scalar",
            #components": [0, 1],# need to add components when changed to vector
            "timeDependentRefData": False,
        },
    },

The total domain of the periodic hill channel is 0<x<9, 0<y<3. So the box I defined is close to the inlet. Then I run the optimization, and this error pops out.

Time = 4000

U Initial residual: (8.771002271424421e-07 3.032438309462439e-07 0.05643213808966887)
U   Final residual: (2.583228439216592e-08 1.638783373336687e-08 0.005531226451734137)
p Initial residual: 6.567763486324534e-06
p   Final residual: 5.277840886562293e-06
Time step continuity errors : sum local = 3.790311999558141e-09
                                 global = -8.967393233392048e-21
                             cumulative = -1.773041426170868e-18
nuTilda Initial residual: 7.884177505128676e-07
nuTilda   Final residual: 4.731480442453503e-08
Bounding nuTilda>1e-16
UFieldVar: 7.279209604676434e-06 final: 7.279209604676434e-06
betaVar: 0.6579903404115919 final: 0.6579903404115919
yPlus min: 0.04722117146807396 max: 0.568779943167962 mean: 0.3408143097966469
ExecutionTime = 123.24 s  ClockTime = 123 s

    Writing cell-centre field C to 4000
    Writing the x component field of the cell-centres Cx to 4000
    Writing the y component field of the cell-centres Cy to 4000
    Writing the z component field of the cell-centres Cz to 4000
End

Printing Primal Residual Statistics.
U Residual Norm2: (3.185704274202537e-07 3.497185123745539e-08 2.397127734052122e-18)
U Residual Mean: (4.266514115025727e-09 2.003633842073476e-10 2.56904504616983e-20)
U Residual Max: (1.277554495938797e-08 6.34085529460475e-09 2.72082254514642e-19)
p Residual Norm2: 2.068401886328395e-06
p Residual Mean: 1.789652285070416e-08
p Residual Max: 5.000025626296046e-07
nuTilda Residual Norm2: 2.419979747523048e-09
nuTilda Residual Mean: 2.873135448567068e-11
nuTilda Residual Max: 1.512689429284541e-10
phi Residual Norm2: 4.553997430715362e-07
phi Residual Mean: 1.62728178838802e-09
phi Residual Max: 1.503339096242937e-08
 
 
Nonlinear constraints
None

Objectives
{'obj.val': array([0.65799762])}

Computing d[UFieldVar]/d[aero_states]^T * psi
Computing d[UFieldVar]/d[beta]^T * psi
Computing d[betaVar]/d[aero_states]^T * psi
[0]PETSC ERROR: ------------------------------------------------------------------------
[0]PETSC ERROR: Caught signal number 11 SEGV: Segmentation Violation, probably memory access out of range
[0]PETSC ERROR: Try option -start_in_debugger or -on_error_attach_debugger
[0]PETSC ERROR: or see https://petsc.org/release/faq/#valgrind
[0]PETSC ERROR: or try http://valgrind.org on GNU/linux and Apple Mac OS X to find memory corruption errors
[0]PETSC ERROR: configure using --with-debugging=yes, recompile, link, and run 
[0]PETSC ERROR: to get more information on the crash.
[0]PETSC ERROR: Run with -malloc_debug to check if memory corruption is causing the crash.
--------------------------------------------------------------------------
MPI_ABORT was invoked on rank 0 in communicator MPI_COMM_WORLD
with errorcode 59.

NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
You may or may not see output from other processes, depending on
exactly when Open MPI kills them.
--------------------------------------------------------------------------

Such error disappears and the code runs perfectly when I use very large boxes, such as min=[-100 -100 -100] max=[100 100 100]. It seems that the box has to be larger than the whole domain in order to avoid this error. Could you give me some hint why this happens? What should I do if I only want to select certain area for my object function? Thanks!

[friedenhe]

We managed to reproduce the issue. There is somehow a seg fault in Computing d[betaVar]/d[aero_states]^T * psi. It happened when the tap.evaluate() is called. We will need to dig into it. We also found that is you add components": [0] to the betaVar function, it can by-pass the error. Not sure why, though...