Dear Professor Ondrej Maxian,
I hope you are doing well.
Thanks for very useful info which you provided in the previous issue.
I have worked on your source for a few months and run the simulations based on my own parameter study. One of these parameters is the "maximum strain".
I have realised that I can lose the accuracy while I run the source code for the different "maximum strain" value. Quite surprisingly, when I repeated the simulation for the same set up, I found that the results are not the exactly the same even when the Brownian motion is deactivated (Note that I have set the "nonLocal=0;" ). This becomes quite noticeable at very small maximum strain value for example, 0.01. I do not know if this is related to compilation or parallelism since the stress for each individual fibre is quite small and this problem can be related to that.
I would be wondering if you could kindly guide me how I can solve this issue and get accurate results.
Best regards,
Ezat
Dear Professor Ondrej Maxian,
I hope you are doing well.
Thanks for very useful info which you provided in the previous issue.
I have worked on your source for a few months and run the simulations based on my own parameter study. One of these parameters is the "maximum strain".
I have realised that I can lose the accuracy while I run the source code for the different "maximum strain" value. Quite surprisingly, when I repeated the simulation for the same set up, I found that the results are not the exactly the same even when the Brownian motion is deactivated (Note that I have set the "nonLocal=0;" ). This becomes quite noticeable at very small maximum strain value for example, 0.01. I do not know if this is related to compilation or parallelism since the stress for each individual fibre is quite small and this problem can be related to that.
I would be wondering if you could kindly guide me how I can solve this issue and get accurate results.
Best regards,
Ezat