Custom Solver Error 1001 with Pulsating 2D TFM fluid bed tutorial

MFiX How to
#1

Hello,

I am using the built-in Pulsating 2D TFM Fluid Bed tutorial with the goal of using similar user defined functions in another project. After successfully building the custom solver with the GUI, I attempted to run the project, but I get the following error:
image

I get the same error if I run through the command line:

image
image772×157 1.49 KB

I am using Fortran compiler 4.8.5 and followed the instructions in Can not build the smp solver under Centos-7 - MFiX - NETL Flow Science Software Support (doe.gov) to remove the “-march=haswell” flag.

Here is my version information:
image

pulsating_fluid_bed_2d_2023-07-13T100849.727410.zip (46.3 KB)

What can I do to fix this?

Thanks,

Ryan

#2

Sorry for asking the obvious question but are you sure the file is present in /uf6/rbreeding/pulsating_fluid_bed_2d and has read permission enabled? I am having trouble coming up with reasons you’d be getting “Unable to open input data file”

#3

It does appear that the file is present and that read permissions are enabled. This issue has also occurred anywhere else I try to use custom solvers. This error does not appear if I use the default solver.
image

#4

GFortran 4.8.5 is quite old and I’m wondering if this might be part of the problem, since you say this affects all custom solvers.

Could you please try this?

$ conda activate mfix-23.2
(mfix-23.2)$ conda install -c conda-forge gfortran

This will install an up-to-date fortran compiler in the Conda environment, without affecting the “system” compiler in /usr/bin.

The current version of GFortran is 13.1.0 and we do not test MFiX with versions older than 8.5.0.

After doing this, go back to your project, click “Clean” and build solver again. (You can also re-enable the Haswell optimizations, which will make the code run faster).

Please let me know if this helps clear up the issue or if further investigation is required.

#5

Although I was unable to update GFortran on that device, I was able to run the solver on another device with GFortran 9.2.0, so it does appear that this was likely the issue. Thanks for the assistance!