File size hasn't been something I've been paying much attention to of late and there are a number of things with how the data sets are saved that can be tweaked to attempt to shrink it (and some things that ended up disabled after tracking down a bug in the HDF5 libraries). I'll take a look shortly (if you could send me one of your files that could be helpful too).
--Daniel
On Jul 5, 2012, at 12:00 PM, Riccardo Pellarin wrote:
> Hi all. I'm using the rmf file support to write coordinates of my
> optimization run.
> I set up the rmf file in this way
>
> rh = RMF.create_rmf_file("traj.rmf")
> IMP.rmf.add_hierarchy(rh, self.prot)
>
> and then at each step "i" I save the coordinates in the file
>
> for i in range(10000):
> mc.optimize(100)
> IMP.rmf.save_frame(rh,i)
>
> Now, the resulting rmf file for my system has a size of about 100 KB per frame,
> which is only slightly lower than the 120 KB per frame of an appended pdb
> file. (My optimization run saves 100000 frames, which makes about 10 GB
> for the rmf). How is it possible that a binary file is that bulky?
> For comparison, a dcd trajectory frame only occupies 18 KB, for the
> same system.
>
> Am I doing something wrong?
>
> cheers
>
> Riccardo
> _______________________________________________
> IMP-dev mailing list
> IMP-dev@salilab.org
> https://salilab.org/mailman/listinfo/imp-dev