24.1.1. custom solver fails and 25.3 cannot enter GUI

MFiX
1

test_2025-11-03T144555.619837.zip (16.0 MB)

(base) amdin@amdin-Rack-Server:~/桌面$ mamba activate mfix-25.3
(mfix-25.3) amdin@amdin-Rack-Server:~/桌面$ mfix
2025-11-03 14:54:44.729 ( 2.708s) [ 7E056C2D5740]vtkOpenGLRenderWindow.c:701 ERR| vtkEGLRenderWindow (0x565e7aee30e0): GLEW could not be initialized: Missing GL version
段错误 (核心已转储)

Dear developer, 24.1.1.always fails to customize solver(without modifying the code), 25.3 cannot enter the GUI after installation.I am using it on the new Ubuntu 22.04 system, and I see that the compiler version is much newer than my previous workstation. I hope to receive your help. Thank you.

2

Based on the error message, it sounds like something I experienced recently too, see link below.

1 Like
3

Kjetil is correct, you can probably fix the issue with 25.3 by doing conda install vtk=*=*qt* with the mfix 25.3 environment activated.

What is the issue you are seeing with 24.1.1?

4

/home/xiejun/miniforge3/envs/mfix-24.1.1/share/mfix/src/model/read_database.f:521:6:

521 | 1010 FORMAT(/8x’Checking Specific Heat is defined along entire temperature range: Yes’)
| 1
Error: Unclassifiable statement at (1)
/home/xiejun/miniforge3/envs/mfix-24.1.1/share/mfix/src/model/read_database.f:595:51:

595 | lMW, HfrefoR(lM,lN)
| 1
Error: FORMAT label 1000 at (1) not defined
/home/xiejun/miniforge3/envs/mfix-24.1.1/share/mfix/src/model/read_database.f:505:61:

505 | Alow(5,lM,lN), Ahigh(5,lM,lN)
| 1
Error: FORMAT label 1020 at (1) not defined
/home/xiejun/miniforge3/envs/mfix-24.1.1/share/mfix/src/model/read_database.f:513:25:

513 | WRITE(ERR_MSG,1010)
| 1
Error: FORMAT label 1010 at (1) not defined
gmake[2]: *** [model/CMakeFiles/mfixcore.dir/build.make:5587model/CMakeFiles/mfixcore.dir/read_database.f.o] 1
gmake[2]: *** …
gmake[1]: *** [CMakeFiles/Makefile2:182model/CMakeFiles/mfixcore.dir/all] 2
gmake: *** [Makefile:136all] 2

                 BUILD FAILED

==========================================================================
the error is displayed as above, aiways read database file and causing compilation issues

5

The VTK issue with version 25.3 has been resolved. Thank you for your help! Version 25.3 allows custom solvers, but 24.1.1 still does not work.

6

Compilation failed mfix-24.4.1: read_database.f:234:44.this question seems a bit like.

7

Yes it’s the same. Can you tell us the version of gfortran on your system?

8

The version of gfortran is 15.2.The gfortran version on my other computer is 13.2,while this 24.1.1 version has been running normally.

9

I uninstalled gfrtran 15,2 and reinstalled version 13.2, which allows me to customize the solver. However, openmai is now missing and dmp is not available. Can you tell me how to correctly lower the version of gfortran?

11

I can verify the problem with gfortran 15.2:

The problem is that in these 2 lines there should be a line after the 8X (the current code has this fixed, but the problem is present in 24.1.1)

1010 FORMAT(/8x'Checking Specific Heat is defined along entire temperature range: Yes')

1020 FORMAT(/8x'Checking Specific Heat is defined along entire temperature range: No', &

This is at line 521 in $CONDA_PREFIX/share/mfix/src/model/read_database.f

You can fix this by opening this file in an editor (outside of MFiX) and inserting a comma between /8x and the quote character:

1010 FORMAT(/8x,'Checking Specific Heat is defined along entire temperature range: Yes')

1020 FORMAT(/8x,'Checking Specific Heat is defined along entire temperature range: No', &

Apparently the older versions of gfortran were more forgiving.

If you fix this, you can go back to using gfortran 15.2.

You may find that you need to reinstall openmpi after changing gfortran versions.

Please let us know if this helps.

– Charles

12

The problems have been solved, and your method is very effective. Thank you for your help！

13

But after setting the particle mass inlet, an error message was displayed, which was not encountered in the CentOS system. The gfortran version is 13.2.Can you give me some solutionsAt line 344 of file des/set_bc_dem_mi.f

Fortran runtime error: Missing comma between descriptors

(2/,2X,'For mass inlet BC: ', I3,/,          4X,'No. particles injected per soli

                                                                                                                                                                                                                                                                               ^



Error termination. Backtrace:

Segmentation fault



Program received signal SIGSEGV: Segmentation fault - invalid memory reference.



Backtrace for this error:

#0 x86_64_fallback_frame_state

  at ./md-unwind-support.h:57

#1 uw_frame_state_for

  at ../../../libgcc/unwind-dw2.c:1016

#2 Unwind_Backtrace

  at ../../../libgcc/unwind.inc:303

#3 backtrace_full

  at ../../../libbacktrace/backtrace.c:127

Segmentation fault
14

This is at line 350 in miniforge3/envs/mfix-24.1.1/share/mfix/src/model/des/set_bc_dem_mi.f，inserting a comma between ES15.8 and ’ PI_COUNT,successfully solved the problem, thank you again for your help

1 Like