Austin hoyle
Student
- Dec 8, 2024
- 3
Hello all,
I am running some pretty computationally demanding jobs on ABAQUS command line. It utilizes a SPH+FEM hybrid method, simulating a penetrator driving into soil. I expected the job to take on the order of 24 hours, since there are many SPH particles to track, so I knew I'd have to run on cmd line instead of in the GUI. As I understand it, if you run a job in the GUI, then ABAQUS will "return the token" after a certain amount of idle time. In practice this does happen, and when I go into the file explorer, all output files stop updating at the time the token is returned, indicating the simulation just freezes in time.
So, my solution was to just run in the cmd line. This was what internet searches suggested. Well I began a simulation yesterday at about 1PM with the "interactive" flag on the command so I can monitor it. When I got back to the office today, the last printed wall time update was 04:41:37, even though it has been about 20 hours since I submitted the job. The cmd window just ends at that. No error message in any output files, the cmd line appears to be waiting for the next print out as if the simulation is still running. Looking in the file explorer, all output files were last updated yesterday about 5 hours after I began the simulation. So all evidence points to the simulation being "frozen".
I thought maybe it was struggling to converge to a solution at this step, but wouldn't there still be output files being updated with new information concerning attempts to converge?
Not entirely sure how to get around this issue. I can't really find anything online suggesting a solution. Hoping someone here maybe is familiar with the problem and knows how to solve it.
It may be helpful to know I am not even close to using all the memory available, nor the storage space on the hard drive.
Thanks in advance!
I am running some pretty computationally demanding jobs on ABAQUS command line. It utilizes a SPH+FEM hybrid method, simulating a penetrator driving into soil. I expected the job to take on the order of 24 hours, since there are many SPH particles to track, so I knew I'd have to run on cmd line instead of in the GUI. As I understand it, if you run a job in the GUI, then ABAQUS will "return the token" after a certain amount of idle time. In practice this does happen, and when I go into the file explorer, all output files stop updating at the time the token is returned, indicating the simulation just freezes in time.
So, my solution was to just run in the cmd line. This was what internet searches suggested. Well I began a simulation yesterday at about 1PM with the "interactive" flag on the command so I can monitor it. When I got back to the office today, the last printed wall time update was 04:41:37, even though it has been about 20 hours since I submitted the job. The cmd window just ends at that. No error message in any output files, the cmd line appears to be waiting for the next print out as if the simulation is still running. Looking in the file explorer, all output files were last updated yesterday about 5 hours after I began the simulation. So all evidence points to the simulation being "frozen".
I thought maybe it was struggling to converge to a solution at this step, but wouldn't there still be output files being updated with new information concerning attempts to converge?
Not entirely sure how to get around this issue. I can't really find anything online suggesting a solution. Hoping someone here maybe is familiar with the problem and knows how to solve it.
It may be helpful to know I am not even close to using all the memory available, nor the storage space on the hard drive.
Thanks in advance!
