New fix ave/chunk and ave running

Dear all,
I wanted to try the new promising chunk/atom commands.
I upgrade from 6 Feb 15 version to the newest 12 Mars 15 and noticed some broken functionality.

First, It seems that a problem appears when averaging chunks using ave running and overwrite options.

As an example, trying to compute a density profile using those commands works:

compute chu Poly chunk/atom bin/1d z lower 0.01 units box
fix out Poly ave/chunk 100 1 100 chu density/number file tmp.out

However trying to time average the results modifying slighltly the command
fix out Poly ave/chunk 100 1 100 chu density/number ave running file tmp.out overwrite

Results in
ERROR: Two fix ave/chunk commands using same compute chunk/atom command in incompatible ways (../compute_chunk_atom.cpp:567)

Despite I only use one fix ave/chunk.
As the size of my simulation box does not change, i tried to fix the problem passing explicitly nchunk once argument in compute. The same error appears when a running average is invoked and disappear when doing one average.

By the way using the previous ave/spatial commands is still working properly as an example:
fix out Poly ave/spatial 100 1 100 z lower 0.01 density/number v_dp v_pn v_pt file Dens.dat ave running overwrite

whith using dp pn and pt as per atom variables.

Is it a Keyboard-Chair interface problem ? Or a more serious issue.
I can provide an example if needed.

The second point is more problematic for me because it breaks backward compatibility.
using rerun command with fix ave/spatial leads to the following error
ERROR: Cannot reset timestep with a time-dependent fix defined (../update.cpp:438)

The same command (same script) used to perfectly works in the 6 Feb Version.
May be a trouble with time depend flag ?



Can you please provide a simple, small input script that

illustrates each of the problems/bugs?

One input script per problem.



Hi Steve,

I send you the minimal examples giving the problems.
The in.chunk_ave show the problem during a running average of chunks trying to compute a density profile.

For the second problem (during rerun), I send you two files, the first in.prep give you a small Dump.gz and data file, needed by the in.rerun.

The in.rerun illustrate the problem found for computing ave/spatial quantity during a rerun.



in.chunk_ave (978 Bytes)

in.prep (547 Bytes)

in.rerun (993 Bytes)