Hello，researchers！After each time I simulate a period of time, Elapsed time and Time remaining (est.) both go up, which were a few days, and then suddenly become tens of thousands of days, but the model keeps going, no divergence, just time to become very large. In addition,I have disabled Enable max wall time. But that doesn’t solve the problem. Why is that? Hope to solve this problem, thank you!
Uploading: zhengchang.png…
Uploading: buzhengchang.png…
Try plotting dt. The time step is computed dynamically and as it gets smaller, the simulation slows down.
But time did not gradually increase, but suddenly increase. Is this also caused by dynamic time steps?
Probably so. Something went wrong with your uploads so I can’t see. Try again?
Go to “main menu”, click “Submit bug report”, upload the ZIP file here
These are pictures before and after the change of time. In addition, relevant documents have also been uploaded. Very much hope that the problem will be solved. Thanks!
1cm-3cm-15cm-0.45mm-30%-water0.013+par-second.mfx (10.7 KB)
I’m missing the particle_input.dat file. Please go to main MFiX menu, click “Submit bug report”, and upload the resulting ZIP file.
Here are the relevant documents, hopefully find out the problem and solve it. Thanks！
bug report.zip (35.5 MB)
In addition, there is a new situation. After the simulation lasts for a while, the time doubles, as shown in the red frame of the picture. But the simulations continue to calculate well. Why is that?
Are you running this in SMP with 99 threads? Have you tried using the serial (single-CPU) solver? How many CPU cores are available on your system?
I did not run this in SMP. My system has a total of 128 CPU cores available, and I runed this in DMP with 96 cores(x=2, y=8, z=6). I have tried using the single-CPU solver, but because the calculation time was too long， I was in the middle of the calculation manually stopped. In trying single-CPU computing this period of time, was no problem.
Ok. There does seem to be something odd going in here with the time-keeping code. I will see if I can replicate this with 2x8x6 DMP. I will let you know what I am able to find. Thanks for your patience.
– Charles
Hope this will be resolved. Looking forward to your response！ Thank you！
I am interested in understanding and fixing this problem but I’m having some trouble reproducing it. Can you give me some more details, such as - how long after starting the simulation did this happen? Does it happen repeatably? Does it happen at the same time in the simulation? Any information you can give me on how to reproduce this would be helpful.
Hello, I will tell you everything I can think of. This happens from time to time. Sometimes this happens a few minutes after starting the simulation, sometimes more than ten hours after starting the simulation, and twenty hours after the simulation starts. But it should be noted that the time after each change is 25620XX, every time, just like shown in the figure. Hope this information will be helpful. Thank you!
Next time this happens, please go to the main menu, click “Enable developer mode”, and open up the “MFiX Status” tab… copy/paste the text in the status window, or take a screenshot, but make sure to capture the whole thing.
Hello, I came across this situation again. The zip file contains the relevant content you may need, hope it helps, thank you！
1.zip (866.0 KB)
Thank you, this is very helpful. This shows me that the problem lies with the solver and not the GUI code.
There are multiple issues related to the timekeeping code in MFiX - we have seen that sometimes CPU time and “wall time” are mixed up. I am planning to review all the time-keeping related code in MFiX. Even though I haven’t been able to reproduce the problem, I think that cleaning up and simplifying the timing code will help with this and several related issues.
This will be fixed for the next MFiX release (23.4 or 23.3.2 if we have another bugfix release). Hopefully this problem is not blocking you completely, it seems like your simulations are running despite this issue.
Thanks for the bug reports!
– Charles
OK. Thank you. I would like to ask if this question affects my simulation results？
In addition, I encountered a new problem during the simulation. In the middle of the simulation, there is a sudden “solver crash” error, as the picture shows, what is the reason? The required mfix file has been uploaded.Thanks！
1cm-3cm-15cm-0.85mm-303.15K-water0.15-20s+10s-second.mfx (12.0 KB)
