HEC-RAS 1D/2D Unsteady Produces Different Error Messages in Different Machines

[AliR]

We have a combined 1D/2D Model and have run it on 2 different machines. The final error messages shown at the end of the model is different in each machine.

We have already cleaned up the model, re-built the geometry and have also run them separately on our local and also in the network drive. We have even installed and uninstalled the HEC-RAS on all machines to ensure no difference. But the results is different in the machines.

We have also tried to run our model on different machines and the results are all similar except 1 machine. The only thing we think might be the difference is the processor of the machine. All other machines that the error results are the same are either 4.0 or 3.0 GHz processor and the machine that the error result is different is 2.2 GHz.

Does the processing power of a machine can affect the HEC-RAS results, or it is something else?
Any suggestions would help.

[Anonymous]

If a model keeps going to maximum iterations than it is probably more likely to have differences between computers.

One thing you should try is go to settings and set it to only use a single core. This turns off parallelization. It will run slower but should be more consistent. I believe this is documented somewhere in the 2D manual.

[AliR]

Thank you for your message.  Yes, some of our runs go to the maximum iterations.  We have re-run our model and have set-up the 2D component of the model  to run with 1 core and following are some of the results (the 1D component was set as default). It definitely helped with the results being more consistent, but as you indicated, it runs slower. 

[Author]

Posts