Hello researchers. I am very curious now: does the MFix-23.3 version support installation on Windows systems? I have repeated this situation six times, and I am not sure what to do. Can you provide some suggestions?
Yes, MFiX 23.3 has been tested on Windows and should install without issues. You should not have to repeat the installation.
The “retrying with next repodata source” message can be ignored.
It’s not clear why the build is failing with no output. Can you please do the following:
Select the “Verbose build output” checkbox and repeat the build. The copy/paste all the text in the build output window (this is better than a screenshot since sometimes the screenshot doesn’t show the full text).
Go to main menu, click “Submit bug report” and upload the resulting ZIP file.
Additionally, an error report has been provided. Could you please help me check where the problem lies? Privately, I also know someone who is using MFix, and they have the same problem on Windows systems. silane_pyrolysis_tfm_2d_2023-10-18T092318.690458.zip (72.5 KB)
I am testing this with MFiX 23.3 on Windows 11 and it works fine. I will try to get my hands on a Windows 10 machine for testing.
There’s really no other output? This is odd. I see you enabled “verbose”. Without any error messages, this is difficult to debug. How about if you try this command at the Anaconda prompt, in your project directory:
Hello, I tried your method and still made the same mistake. I always encounter this error during the installation process; I don’t know why. Is it due to gfortran or cmake?
Note that after the message about current_repodata.json. it retries and succeeds - the environment is successfully solved and packages are downloaded. This is a normal, successful installation. If you want to know more about the current_repodata message and why it appears, please read this section of the Conda documentation, although understanding Conda at this level is not necessary to install and use MFiX.
This problem also has nothing to do with the failure to build the solver. Installation and solver compilation are two completely separate processes. Let’s concentrate on the solver build failure.
Now, what output do you get when you run these commands?
Regarding the installation warnings: You may also be interested in using Mamba to install MFiX - this runs much faster than conda, and produces fewer warning messages.
Ok. I have entered the command three times. Strangely, the first command did not respond; was the second command correct? The third command displays an error message.
Additionally, I input commands on the control end and found that cmake has been successfully installed. However, entering the command in (mfix 23.3.1) did not result in any response, and I don’t know why.
This is very strange! I am trying to find an answer, but this is really not an issue with MFiX at all but with the Conda system. I’ll let you know what I find.
Thank you very much for your reply. I will always wait. But due to the time difference, I may not be able to reply to you in a timely manner. I’m very sorry.
No worries, we have users all around the world in different time zones, and nobody expects immediate replies (or at least, they shouldn’t!). Also note I will be on vacation all next week, so you won’t hear from me for a little while.
I suppose one thing you could try is forcing a different CMake version:
# Show currently installed verion
(mfix-23.3.1)C:> conda list cmake
# packages in environment at C:\Users\cgw\mambaforge\envs\mfix-23.3.1:
#
# Name Version Build Channel
cmake 3.27.6 hf0feee3_0 conda-forge
# Shows all available versions
(mfix-23.3.1)C:> conda search -c conda-forge cmake
# Name Version Build Channel
cmake 3.5.0 0 conda-forge
...
cmake 3.27.5 hf0feee3_0 conda-forge
cmake 3.27.6 hf0feee3_0 conda-forge
MFiX requires CMake version at least 3.16 (or maybe 3.24?) but you could try installing some different CMake versions and see if it helps:
Keep trying different versions, working backward until you find one that works (produces output when you run cmake --version).
If this works for you please let us know!
It looks like there may be a problem with the cmake 3.27.4 in conda-forge (the library MSVCP140D is a debug library, production code should not depend on this, so perhaps this package was improperly built using debug mode). Version 3.27.6 is reported to work. So conda install -c conda-forge cmake=3.27.6 should fix the issue for you.