Sure, the Linux Anacond version of MFiX should work.
If it is Ubuntu based then you will be able to:
apt install gfortran
for building the solver, and
apt install libopenmpi-dev
if you want to build with MPI.
thanks you a lot mark, im mechanical engineer from Chile and i use OpenFOAM and dpm for mining aplications in the freelance way.
I hope to test the code more deeply and show it for Latin America, I am amazed how the gui has improved in recent years.
Thanks!
Glad to hear it! Hope you continue to enjoy using MFiX.
I use the parallel in window. Open anaconda>>“activate mfix-19.3.1”>> cd to your directory >> type build_mfixsolver --smp to build the solver in that directory>> then type #!MFIX-GUI OMP_NUM_THREADS = number of threads you want to use at the mfix file __.mfx or mfix.dat After that you can run the mfix with parallel
What directory is it? can you made some screenshoots
2 Likes
Sorry for slowly respomse, I have just see the noti in email
CD is the command which uses to change the working directory. For this case, working directory need to point for creating the new custom server. Don’t worry, Please feel free for further discussion.
Thank you for your kind reply! Actually in the morning I just knew the meaning of cd and I thought it was a little awkward so I delet my question… Now I have another question. Follow your guide, I can use the parallel in windows. But I don’t know how to restart after I stop the calculation. Can you help me?
In graphical interface, after we click reset button (button1). The MFIX will ask to delete the file with that extension (.RES, .SP, .vtp,.vtu,.pvd, .csv, .Out, .pid, … similar to picture) before restart the job. By the way, if you run the MFIX in command interface. You need to delete that extension in folder before restart the job.
Don’t worry.
There’s a first time for everyone. Enjoy to use MFIX
Thanks! But I am so sorry that I use the wrong word. I try to resume, not restart.
When running the mfix in command interface, the program works in command interface, rather than the graphical interface, like this picture.
Actually the only way I found to stop calculation is clicking the botttom line in command interface. And I press Enter to continue the program. At noon I stopped the calculation and went out for a meal, then I found that I don’t know how to proceed with the calculation. Finally I had to recalculate from scratch.
From your advice, I know I need to delete these files before restart. Now I very much hope you can give me advice about resuming. For example, what statement should I enter into the command interface?
-You can learn further command through this user manual https://mfix.netl.doe.gov/doc/mfix/latest/pdf/MFiX-User_Guide.pdf.
Thanks! I will have a try.
Hi! I try this method but after the step 5 the solver starts without asking the number of threads. I also used the GUI but the SMP option continue disabled. I will really appreciate your help! Thank you in advance!
Please capture the anaconda command at step 5, I can guide you somehow.
Hi! I done the procedure several times changing folders and simulations, and it works when I restart the Anaconda prompt after the solver creation. Now I can choose the number of threads (both in command and GUI) but with any number different of 1 the log says "Process failed to start env OMP_NUM_THREADS=2 "
Hi, Mark,
Can parallel computing of mfix case be performed on a virtual machine installed in WIN7 operating system? Compare to that run on Linux system directly, to what extent the calculating speed will be affected?
Yes, using Linux within a virtual machine such as VirtualBox or VMWare on Windows would works. Using WSL2 would also be possible. The performance for a laptop or desktop system in a virtual environment would likely be comparable to running Linux natively on the same hardware (I would guess.)
Hi! I am doing a CFD-DEM simulation of a cylindrical liquid-solid fluidized bed. I have tried the method that @tyurata described and indeed it works, but only with the Hookean collision method. If I try to change it to Hertzian, the program computes one time step, and then it stops without giving any error. Does anybody know if this can be solved? Thanks!
Hi Filippo - can you upload your project files, and we’ll have a look? You can use ‘Submit bug report’ to create a zip file containing everything.
- Charles