Re: [AMBER] cpptraj frames clustering limit

From: Daniel Roe <>
Date: Wed, 12 Aug 2015 09:06:15 -0600

On Wed, Aug 12, 2015 at 9:46 AM, Gaurao Dhoke <> wrote:
> I have found that cpptraj is not able to process long trajectory for
> analysis.

It would help if you could provide more information on what you mean
here. Does cpptraj print a warning/error message? Does it crash?

On Wed, Aug 12, 2015 at 7:59 AM, Jason Swails <> wrote:
> On Wed, Aug 12, 2015 at 9:46 AM, Gaurao Dhoke <> wrote:
>> Is there any way to increase this limit?
> Since you're doing clustering, either your system is running out of memory
> or you are expecting a calculation that may take hours or days to finish
> much faster than that (so you think it's hanging, when in fact it's still
> working). You haven't provided enough information about what happened for
> me to know for sure.
> If you are actually running out of memory, your choices are either use a
> machine that has more RAM, or use the "sieve" option, which will generate
> the clusters based on a reduced number of frames, and then assign *all* of
> the frames to that cluster. More details are available in the manual.

There is actually a third option - you can use the 'loadtraj' command
to load your frames as a trajectory (TRAJ) data set, which functions
like a coordinates (COORDS) data set except everything is kept
on-disk. The 'loadtraj' command can be used to make a TRAJ data set
out of existing 'trajin' entries, or can be used to load a trajectory
separately. The only limitation is that all trajectories in a TRAJ
data set must correspond to the same topology file. So you could do
for example:

parm myparm.parm7
loadtraj name crd1
cluster crdset crd1 <more cluster options>

See the Amber 15 manual section 29.4 for more on using coordinates as
data sets in cpptraj. Note that you will still need enough memory to
store the pairwise distance matrix, so you may need to use the 'sieve'
option as well depending on how much memory you have available.

Hope this helps,


> Also, I would highly recommend upgrading to AmberTools 15 if you haven't
> already (if you have Amber 14, you can do this with a simple
> "./update_amber --upgrade" command).
> HTH,
> Jason
> --
> Jason M. Swails
> BioMaPS,
> Rutgers University
> Postdoctoral Researcher
> _______________________________________________
> AMBER mailing list

Daniel R. Roe, PhD
Department of Medicinal Chemistry
University of Utah
30 South 2000 East, Room 307
Salt Lake City, UT 84112-5820
(801) 587-9652
(801) 585-6208 (Fax)
AMBER mailing list
Received on Wed Aug 12 2015 - 08:30:05 PDT
Custom Search