Hello.
My MD run is 600ns. This run finished in an hour.
I used the following command for clustering:
gmx trjconv -f md.xtc -s md.tpr -o md_cluster.gro -e 600000 -pbc cluster
After a day, it has not been finishing! I was wondering if it is normal.
I used 600000 as time of last frame. Is it correct?
Hi,
The issue is the -cluster option here that can take a long time. It’s depending on what you want to analyse, but usually it’s helpful to run a first trjconv round where you reduce the amount of data first. For most analysis trjconv -pbc whole is also a good choice and should not take that long.
Another reason for things to take a long time might be a remote connection, e.g., if you have a directory mounted via sshfs and run GROMACS on your local computer for analysis, though in your case it’s most likely the -cluster
Thank you so much for your answer. It is a micellar structure and I want to calculate the radius of gyration, so I think I have to chose -cluster.
Which value should I choose to reduce the amount of data? -round (yes/0/1/true/false)
Hi,
Oh, sorry this was confusingly phrased - I did not mean that you run trjconv -round, but that you run another round of trjconv.
Usually it’s better to reduce the amount of data by choosing
only the part of the system that you want to analyse with -index - note that for this, structure files will also have to match which you can achieve by gmx convert-tpr -index, generating a new tpr file that will help you with analysis only
skipping over time-frames in your trajectory using -skip - this is the easier and more promising option in your case; this will allow you to see if your approach works, then you can go to smaller -skip values and run a longer conversion run, after you confirmed with a large -skip value (something around 100 or 1000, depending on your system and input trajectory) that things go as you intend
Thank you so much.