Skip to content

DCD info: trajectory.delta wrong, trajectory.dt in wrong units #84

@GoogleCodeExporter

Description

@GoogleCodeExporter
What steps will reproduce the problem?
1. u = Universe("brown.psf", "brown.dcd")
2. print u.trajectory.delta
   9.99999974738e-05
3. print u.trajectory.dt
  0.0048888208765
4. print u.trajectory.units["time"]
  AKMA
5.print 
u.trajectory.dt*MDAnalysis.core.units.timeUnit_factor[u.trajectory.units["time"]
]/u.trajectory.skip_timestep
  0.0999999974738

What is the expected output? What do you see instead?

for 2:
Expected: 0.0001
Instead: 9.99999974738e-05
Relative Error: 2.52621248364e-08

for 3:
Expected: 0.1
Instead: 0.0048888208765

for 4:
I do not know what I should expect. No units?

for 5:
Expected: 0.1
Instead: 0.0999999974738
Relative Error: 2.52621247254e-08

What version of the product are you using? On what operating system?
MDAnalysis 0.7.5 and MDAnalysis 0.7.4
Linux ubuntu 11.04 AMD 64 2.6.38-13
python 2.7





Please provide any additional information below.

Concerning 4:
The .lam file is in lj units, so the output should be in dimensionless
time units.
Here, the system seems to believes that the data is in AKMA units and converts 
internally to some other time unit.
But even when only considering the values for .delta, I still have a relative 
error of 2.e-8 even for simple values such as 0.0001.
Note that the error is zero for values of delta = 1.

This error is a pain in the ass, for long timeseries this adds up to more than 
one frame.



Original issue reported on code.google.com by bjrnfrd...@gmail.com on 14 Nov 2011 at 3:36

Attachments:

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions