Set up a trajectory ensemble for reading during a run.

ensemble <file0> {[<start>] [<stop> | last] [offset]} | lastframe [parm <parmfile / tag> | parmindex <#>] [trajnames <file1>,<file2>,...,<fileN> [remlog <remlogfile> [nstlim <nstlim> ntwx <ntwx>]]

<file0> Lowest replica filename.
[<start>] Frame to begin reading ensemble at (default 1).
[<stop> | last] Frame to stop reading ensemble at; if not specified or ’last’ specified, end of trajectories.
[<offset>] Offset for reading in trajectory frames (default 1).
[lastframe] Select only the final frame of the trajectories.
[parm <parmfile>] Topology filename/tag to associate with trajectories (default first topology).
[parmindex <#>] Index of Topology to associate with trajectories (default 0, first topology).
[trajnames <file1>,...,<fileN>] Do not automatically search for additional replica trajectories; use comma-separated list of trajectory names.
[remlog <remlogfile>] For H-REMD trajectories only, use specified REMD log file to sort trajectories by coordinate index (instead of by Hamiltonian)[nstlim <nstlim> ntwx <ntwx>] If trajectory and REMD log were not written at the same rate, these are the values for nstlim (steps between each
exchange) and ntwx (steps between trajectory write) used in the REMD simulation.

Read in and process trajectories as an ensemble. Similar to ’trajin remdtraj’, except instead of processing one frame at a target temperature, process all frames. This means that action and trajout commands apply to the entire ensemble; note however that not all actions currently function in ’ensemble’ mode. For example, to read in a replica ensemble, convert it to temperature trajectories, and calculate a distance at each temperature:

parm ala2.99sb.mbondi2.parm7
ensemble rem.crd.000 trajnames rem.crd.001,rem.crd.002,rem.crd.003
trajout temp.crd
distance d1 out d1.ensemble.dat @1 @21

This will output 4 temperature trajectories named ’temp.crd.X’, where X ranges from 0 to 3 with 0 corresponding to the lowest temperature, and ’d1.ensemble.dat’ containing 4 columns, each corresponding to a temperature. If run with MPI, data will be written to separate files named ’d1.ensemble.dat.X’, similar to the output trajectories.

Note that in parallel (i.e. MPI) users should specify the ensemblesize command prior to ensemble in order to improve set up efficiency.