Skip to content
Snippets Groups Projects

[Flow-v1.3] API change and improved scripts

Merged Boman Romain requested to merge adrien into master

Created by: acrovato

Context

This PR brings the latest changes in Flow to master. This version is the minimal requested to run CUPyDO-v1.0

Modifications

  • Split Solver class into Newton and Picard classes, user can choose the algorithm as they see fit
  • Improved polar scripts, the solver and the problem are not re-initialized at every change in angle of attack
  • Created template functions in python calling the C++ classes and functions used in the tests, if the C++ code changes, then there is only one python file to adapt
  • Added CODEOWNERS file
  • Minor cleaning/refactoring
  • Minor refactoring by @rboman

Tests

Tests are passing on debian-4.9 (gaston), ubuntu-16.04 (shippable, no Trilinos) and ubuntu-18.04 (my VM)

Comments

  • Regarding the template functions I created for flow tests and validation (flow/template.py): I am not sure about this new method, I will probably refactor it into real python API in the future
  • Regarding the CODEOWNERS file: this is a test, I will try to PR without adding any reviewers

Merge request reports

Approval is optional

Merged by avatar (Apr 15, 2025 3:26am UTC)

Merge details

  • Changes merged into master with cc047caa.
  • Deleted the source branch.

Activity

Filter activity
  • Approvals
  • Assignees & reviewers
  • Comments (from bots)
  • Comments (from users)
  • Commits & branches
  • Edits
  • Labels
  • Lock status
  • Mentions
  • Merge request status
  • Tracking
  • Created by: acrovato

    Once this PR is completed, we can proceed with the cleaning of tab/spaces

  • Created by: acrovato

    The CODEOWNERS file did not trigger anything (it should have added @rboman ). I have noticed a "codeowner" option in the repo settings. I enabled it, but I guess we should make a new PR to check ;P

  • The CODEOWNERS file did not trigger anything (it should have added @rboman ). I have noticed a "codeowner" option in the repo settings. I enabled it, but I guess we should make a new PR to check ;P

    I guess it takes into account the CODEOWNERS file of the master branch

    I'm compiling on my machines...

  • 00abf1e354ff7b6a343204cabb075e6c30bf69292aae997d604e0c6c0fbe325f

    92% tests passed, 5 tests failed out of 60
    
    Total Test time (real) =  41.47 sec
    
    The following tests FAILED:
    	  6 - tbox/tests/meshDeformation3.py (Failed)
    	 16 - flow/tests/cylinder2D5.py (SEGFAULT)
    	 17 - flow/tests/cylinder3.py (Failed)
    	 19 - flow/tests/lift3.py (Failed)
    	 21 - flow/tests/meshDef3.py (Failed)
    Errors while running CTest
    rboman@spirou:~/dev/Projet_MP/waves/build$ 
  • Created by: acrovato

    The following tests FAILED:
    	  6 - tbox/tests/meshDeformation3.py (Failed)
    	 16 - flow/tests/cylinder2D5.py (SEGFAULT)
    	 17 - flow/tests/cylinder3.py (Failed)
    	 19 - flow/tests/lift3.py (Failed)
    	 21 - flow/tests/meshDef3.py (Failed)
    Errors while running CTest

    Meeeeeh. Let me guess: singular matrix in mumps?

  • Boman Romain
    Boman Romain @rboman started a thread on commit cc047caa
1 #!/bin/bash
2 # SLURM script for WAVES ctest
3 #SBATCH --job-name=waves
4 #SBATCH --mail-user=boman
5 #SBATCH --mail-type=ALL
6 # Ressources needed...
7 #SBATCH --partition=defq
8 #SBATCH --ntasks=1
9 #SBATCH --cpus-per-task=12
10 #SBATCH --time=0-1:00:00
11 #SBATCH --mem=10000
12 . /home/boman/.bash_profile fabulous
13 cd build
14 pwd
15 echo "running ctest..."
16 srun ctest -j 12
  • Boman Romain
    Boman Romain @rboman started a thread on commit cc047caa
  • 1 #!/bin/bash
    2 # SLURM script for WAVES ctest
    3 #SBATCH --job-name=waves
    4 #SBATCH --mail-user=boman
    5 #SBATCH --mail-type=ALL
    6 # Ressources needed...
    7 #SBATCH --partition=defq
    8 #SBATCH --ntasks=1
    9 #SBATCH --cpus-per-task=12
    10 #SBATCH --time=0-1:00:00
    11 #SBATCH --mem=10000
    12 . /home/boman/.bash_profile fabulous
    13 cd build
    14 pwd
    15 echo "running ctest..."
    16 srun ctest -j 12
  • Boman Romain
    Boman Romain @rboman started a thread on commit cc047caa
  • 1 #!/bin/bash
    2 # SLURM script for WAVES ctest
    3 #SBATCH --job-name=waves
    4 #SBATCH --mail-user=boman
    5 #SBATCH --mail-type=ALL
    6 # Ressources needed...
    7 #SBATCH --partition=defq
    8 #SBATCH --ntasks=1
    9 #SBATCH --cpus-per-task=12
    10 #SBATCH --time=0-1:00:00
    11 #SBATCH --mem=10000
    12 . /home/boman/.bash_profile fabulous
    13 cd build
    14 pwd
    15 echo "running ctest..."
    16 srun ctest -j 12
  • Meeeeeh. Let me guess: singular matrix in mumps?

    no, bad gmsh - let me check with a newer one... (my macos will do an upgrade tonight, so more errors could appear tomorrow) :p

  • Created by: acrovato

    Meeeeeh. Let me guess: singular matrix in mumps?

    no, bad gmsh - let me check with a newer one... (my macos will do an upgrade tonight, so more errors could appear tomorrow) :p

    3.0.6 is your best pick. tests/nonlift.py does not fully converge with 4.x.x. (but the test still passes).

  • steve-jobs-6_1980157c

    rboman@spirou:~/dev/Projet_MP/waves/build$ ctest -j 6 --rerun-failed
    Test project /Users/rboman/dev/Projet_MP/waves/build
        Start  6: tbox/tests/meshDeformation3.py
        Start 16: flow/tests/cylinder2D5.py
        Start 17: flow/tests/cylinder3.py
        Start 19: flow/tests/lift3.py
        Start 21: flow/tests/meshDef3.py
    1/5 Test  #6: tbox/tests/meshDeformation3.py ...   Passed    1.03 sec
    2/5 Test #21: flow/tests/meshDef3.py ...........   Passed   11.31 sec
    3/5 Test #17: flow/tests/cylinder3.py ..........   Passed   13.26 sec
    4/5 Test #16: flow/tests/cylinder2D5.py ........   Passed   14.38 sec
    5/5 Test #19: flow/tests/lift3.py ..............   Passed   23.80 sec
    
    100% tests passed, 0 tests failed out of 5
    
    Total Test time (real) =  23.81 sec
    rboman@spirou:~/dev/Projet_MP/waves/build$ 
    `
    ``
  • The battery is successful on macOS! Let's celebrate... before tomorrow: I have delayed a planned update

  • Please register or sign in to reply
    Loading