lindahl at csb.stanford.edu
Thu Oct 16 09:19:25 CEST 2003
> Are there any big differences to be expected between the current binary
> format used for xtc,... and netcdf ? What eg about compression ? I
> know much about netcdf I must admit, but for now, the current
> fulfills all my needs and seems very efficient.
NetCDF can't do anything like xtc compression since we use a lot of
knowledge about molecular system to improve the lossy compression, but
all the other file formats just write uncompressed data.
Whatever we do, the current formats will still be there, but we might
add a new one.
I've been playing with the idea of having a file format with a header
containing a simple topology of the system, including bonds, atomtypes
& names, and charges. If it is only present in the first frame it won't
increase the size significantly, but with a standard format you could
forget about keeping those tpr files around for analysis or viewing.
More information about the gromacs.org_gmx-developers