On issues about xyz format of plumed driver

[Hu Li]

Hi, friends.

Currently, I am using driver to analysis my trajectories which is biased by metaD using Structure Factor. However, it did not go smothingly. 

I found the data calculated by driver are all 0s and nans when cutoff is 4.0, using exactly the same trajectory and the same plumed.dat files.  If I changed cutoff to 6.0, then it can give numbers but show inconsistence to on-the-fly computaion, I shared picture below.

I suspect it was because of the format. The xyz format were changed to extremely clean-out one from GPUMD, because plumed can't cope with the original on. However, I am not sure if this is what pluemd wanted, but it can goes well on other cvs except structure factor.  

I eagerly need some one to share the standard format of xyz file which plumed can treat it well. And if possible, may anyone can explain the inconsistency calculated by on-the-fly computation and the subsequent post-processed computation？ 

Appreciate it ！

Weihu Li

[Hu Li]

Dear all:

Sorry to bother again, this is my answer as a mark for my dumb mistake. I have always known the unit in Plumed is nm, so I added units equal to angstrom to plumed.dat from beginning. However, I just found that it would  only  work when you added it to the executive command. plumed driver --ixyz clean_output1.xyz --plumed plumed.dat --length-units A

So my xyz_file is ok here, and the problem is solved simply because the mismatch of units, thanks and sorry to occupied precious resources.
Weihu