mdolab/MACH-Aero

Bug when running MACH-Aero tutorial with SLSQP

1064168551 opened this issue · 8 comments

1064168551 commented

Description

There is a bug when running MACH-Aero tutorial using docker named mdolab/public:u20-gcc-ompi-stable, I would be very grateful if you could help me.
The bug is

'ValueError: failed to initialize intent(inout) array -- expected elsize=8 but got 16 -- input 'D' not compatible to 'd'
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
  File "/home/mdolabuser/mount/opt/aero/aero_opt.py", line 245, in <module>
    sol = opt(optProb, MP.sens, storeHistory=os.path.join(args.output, "opt.hst"))
  File "/home/mdolabuser/.pyenv/versions/3.9.16/lib/python3.9/site-packages/pyoptsparse/pySLSQP/pySLSQP.py", line 218, in __call__
    slsqp.slsqp(m, meq, la, n, xs, blx, bux, ff, gg, df, dg, acc, maxit,
ValueError: failed in converting 5th argument `x' of slsqp.slsqp to C/Fortran array'

Steps to reproduce issue

  1. …
  2. …
  3. …

Current behavior

…

Expected behavior

…

Code versions

  • Operating System:
  • Python:
  • OpenMPI:
  • CGNS:
  • PETSc:
  • Compiler:
  • This repository:
1064168551 commented

The figure of the bug was listed, I would be really grateful if you can help me.

eirikurj commented

Thanks for the issue. For completeness, please update and populate the remaining issue sections, i.e., Steps to reproduce issue with minimal steps etc.

However, I can confirm that the tutorial is broken when running SLSQP. This is due to a recent change in pyOptSparse (see PR mdolab/pyoptsparse#348 for details). We will look into this, but the fundamental problem is that pyGeo DVs are allocated as complex.

In the meantime to circumvent the issue, you can checkout the previous version of pyOptSparse into a separate branch, e.g., git checkout -b v2.10.0 v2.10.0.

1064168551 commented

Sincerely thanks for your answer and kindness, SLSQP seems OK when I used pyOptSparse 2.10.0.

A-CGray commented

For future reference, the solution here is to fix the handling of real and complex values in pyGeo

r-quadros commented

@eirikurj : I am running the optimization in docker and facing the same challenge. I am trying to revert back to a previous version of pyOptSparse still the error remains. Here is where I am reverting back in the "repos" folder in the container as seen in the "git status":
mdolabuser@bf5736a488a8:/repos/pyoptsparse$ git status
On branch v2.10.0
nothing to commit, working tree clean
mdolabuser@bf5736a488a8:/repos/pyoptsparse$

and when I try to run the optimization, I get the same error. Is there something I am missing here? I am sorry, I am new to docker and probably I am reverting to a previous branch in a wrong folder.

A-CGray commented

Did you pip install pyoptsparse again after checking out the older release?

r-quadros commented

@A-CGray : That helped immensely! Opt. running now. Many thanks.