Logic error in TimeStep Control is triggered after 1st iteration of 3rd NS step:

[jcschmitt]

This is a minimal example demonstrating a bug.

Input VMEC file is attached.
input.vmec.txt

[jcschmitt]

Here is the output log of the VMEC run with the file attached.

 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
  SEQ =    1 TIME SLICE  0.0000E+00
  PROCESSING INPUT.vmec
  THIS IS PARVMEC (PARALLEL VMEC), VERSION 9.0
  Lambda: Full Radial Mesh. L-Force: hybrid full/half.

  COMPUTER: linux1   OS: GNU/Linux   RELEASE: 6.5.0-18-generic  DATE = Mar 25,2024  TIME = 18:49:35

  NS =   16 NO. FOURIER MODES =  128 FTOLV =  1.000E-14 NITER =   2000
  PROCESSOR COUNT - RADIAL:    2
 INITIAL JACOBIAN CHANGED SIGN!
 TRYING TO IMPROVE INITIAL MAGNETIC AXIS GUESS
  ---- Improved AXIS Guess ----
      RAXIS_CC =    8.40000000000000      -0.784375943362629     
 -1.901116593816534E-016 -3.061680294433361E-015 -2.512147933894040E-015
 -1.452335524282492E-015 -7.064929488888176E-015  5.966351342998344E-015
 -9.987984650384246E-015
      ZAXIS_CS =   0.000000000000000E+000  0.403265727856953     
  3.403911385037090E-017  1.747954495019046E-017 -1.112179723293405E-016
 -2.250874735691193E-016 -1.882884315687183E-016  4.072427314711041E-016
 -9.560353638696670E-017
  -----------------------------

  ITER    FSQR      FSQZ      FSQL    RAX(v=0)    DELT       WMHD

    1  4.49E-03  7.58E-04  4.13E-03  7.616E+00  7.00E-01  1.1500E+04
  200  4.48E-08  2.72E-09  2.51E-08  7.685E+00  7.00E-01  1.1435E+04
  400  2.62E-08  5.83E-09  1.67E-08  7.903E+00  7.00E-01  1.1433E+04
  600  2.29E-08  5.41E-09  9.60E-09  8.029E+00  7.00E-01  1.1432E+04
  800  2.02E-08  4.37E-09  6.06E-09  8.057E+00  7.00E-01  1.1432E+04
 1000  1.33E-08  3.09E-09  3.71E-09  8.075E+00  7.00E-01  1.1431E+04
 1200  4.43E-09  1.19E-09  1.25E-09  8.087E+00  7.00E-01  1.1431E+04
 1400  2.28E-09  5.28E-10  3.63E-10  8.093E+00  7.00E-01  1.1431E+04
 1600  8.20E-10  2.06E-10  1.86E-10  8.095E+00  7.00E-01  1.1431E+04
 1800  4.27E-10  1.18E-10  1.24E-10  8.097E+00  7.00E-01  1.1431E+04
 2000  2.71E-10  9.70E-11  7.87E-11  8.098E+00  7.00E-01  1.1431E+04

  NS =   32 NO. FOURIER MODES =  128 FTOLV =  1.000E-13 NITER =   6000
  PROCESSOR COUNT - RADIAL:    2

  ITER    FSQR      FSQZ      FSQL    RAX(v=0)    DELT       WMHD

    1  5.63E+00  4.95E+00  1.28E-02  8.098E+00  7.00E-01  1.1433E+04
  200  2.50E-06  1.20E-06  2.19E-08  8.108E+00  7.00E-01  1.1431E+04
  400  1.94E-07  7.47E-08  6.16E-09  8.127E+00  7.00E-01  1.1431E+04
  600  4.98E-08  1.51E-08  4.04E-09  8.142E+00  7.00E-01  1.1431E+04
  800  1.55E-08  5.24E-09  2.75E-09  8.147E+00  7.00E-01  1.1431E+04
 1000  8.22E-09  2.94E-09  2.32E-09  8.149E+00  7.00E-01  1.1431E+04
 1200  6.68E-09  2.06E-09  1.90E-09  8.151E+00  7.00E-01  1.1431E+04
 1400  4.06E-09  1.22E-09  1.60E-09  8.154E+00  7.00E-01  1.1431E+04
 1600  3.19E-09  9.15E-10  1.47E-09  8.155E+00  7.00E-01  1.1431E+04
 1800  3.00E-09  7.74E-10  1.34E-09  8.157E+00  7.00E-01  1.1431E+04
 2000  2.83E-09  7.62E-10  1.15E-09  8.158E+00  7.00E-01  1.1431E+04
 2200  2.99E-09  8.02E-10  9.71E-10  8.159E+00  7.00E-01  1.1431E+04
 2400  3.04E-09  8.36E-10  8.37E-10  8.160E+00  7.00E-01  1.1431E+04
 2600  3.07E-09  8.84E-10  7.31E-10  8.161E+00  7.00E-01  1.1431E+04
 2800  3.11E-09  9.42E-10  6.53E-10  8.162E+00  7.00E-01  1.1431E+04
 3000  3.11E-09  9.61E-10  5.88E-10  8.163E+00  7.00E-01  1.1431E+04
 3200  3.08E-09  9.10E-10  5.27E-10  8.164E+00  7.00E-01  1.1431E+04
 3400  3.08E-09  8.18E-10  4.78E-10  8.165E+00  7.00E-01  1.1431E+04
 3600  2.87E-09  6.90E-10  4.48E-10  8.165E+00  7.00E-01  1.1431E+04
 3800  2.47E-09  5.67E-10  4.19E-10  8.166E+00  7.00E-01  1.1431E+04
 4000  1.98E-09  4.60E-10  3.84E-10  8.166E+00  7.00E-01  1.1431E+04
 4200  1.54E-09  3.66E-10  3.49E-10  8.166E+00  7.00E-01  1.1431E+04
 4400  1.17E-09  2.93E-10  3.14E-10  8.167E+00  7.00E-01  1.1430E+04
 4600  8.70E-10  2.25E-10  2.78E-10  8.167E+00  7.00E-01  1.1430E+04
 4800  6.45E-10  1.70E-10  2.43E-10  8.167E+00  7.00E-01  1.1430E+04
 5000  4.69E-10  1.23E-10  2.10E-10  8.167E+00  7.00E-01  1.1430E+04
 5200  3.52E-10  8.74E-11  1.83E-10  8.168E+00  7.00E-01  1.1430E+04
 5400  2.71E-10  6.35E-11  1.60E-10  8.168E+00  7.00E-01  1.1430E+04
 5600  2.09E-10  4.76E-11  1.39E-10  8.168E+00  7.00E-01  1.1430E+04
 5800  1.59E-10  3.67E-11  1.19E-10  8.168E+00  7.00E-01  1.1430E+04
 6000  1.24E-10  2.84E-11  9.99E-11  8.168E+00  7.00E-01  1.1430E+04

  NS =   64 NO. FOURIER MODES =  128 FTOLV =  1.000E-12 NITER =   8000
  PROCESSOR COUNT - RADIAL:    2
 INITIAL JACOBIAN CHANGED SIGN!
 TRYING TO IMPROVE INITIAL MAGNETIC AXIS GUESS
  ---- Improved AXIS Guess ----
      RAXIS_CC =    8.75043008571858      -0.754691882966337     
 -0.104038015487500      -2.190653195299761E-003 -7.089092641349942E-004
 -3.772769326218403E-003 -4.181896415598869E-002  6.363489331780362E-003
  1.782947574331021E-002
      ZAXIS_CS =   0.000000000000000E+000  0.255895057905578     
  5.848548619850485E-003 -6.473203901076727E-003  8.395084630594362E-004
 -1.560975077431877E-002  1.677437583572644E-004  1.119731424046128E-002
 -1.583869981538949E-004
  -----------------------------

  ITER    FSQR      FSQZ      FSQL    RAX(v=0)    DELT       WMHD

    1  1.24E-10  2.84E-11  9.99E-11  0.000E+00  7.00E-01  0.0000E+00
Logic error in TimeStepControl!
Logic error in TimeStepControl!

===================================================================================
=   BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
=   RANK 0 PID 571151 RUNNING AT linux1
=   KILLED BY SIGNAL: 9 (Killed)
===================================================================================

[cianciosa]

Two things I noticed is R axis and the MHD energy is zero. It looks like it didn't reset the magnetic axis correctly.

[cianciosa]

I turned off PARVMEC to run the SERIAL versions of the subroutines. On the 3rd grid, it fails to find a good initial axis guess for the axis 3 times then terminates. So even if I fix the logic control so the PARALLEL and SERIAL match, this will still have problems with this case.