mivade / ionmd Goto Github PK

Params._pack_ = 4 should read Params._pack_ = 8 on 64 bit CPUs. This is either a bug in h2xml or xml2py, though there may be an option passable to one or both to fix this. As far as I know, this is only a problem when CCD simulation is enabled.

Need to make it so that matplotlib/maya/PIL are truly optional (exception handling, abort functions if they aren't around, etc.).

CMake output indicates that it works:

-- Try OpenMP C flag = [-fopenmp]
-- Performing Test OpenMP_FLAG_DETECTED
-- Performing Test OpenMP_FLAG_DETECTED - Success
-- Try OpenMP CXX flag = [-fopenmp]
-- Performing Test OpenMP_FLAG_DETECTED
-- Performing Test OpenMP_FLAG_DETECTED - Success
-- Found OpenMP: -fopenmp  
-- Configuring done

... but then building results in the warning

ionmd/ionmd.cpp:48:0: warning: ignoring #pragma omp parallel [-Wunknown-pragmas]
     #pragma omp parallel for
 ^

As more is being added, the main directory is getting rather messy. I propose changing the structure to something like this:

IonMD
|-- src
|    `-- *.cpp
|    `-- *.hpp
|-- doc
|    `-- <documentation files other than README>
|-- data
|-- images
`-- <python files>
`-- README.rst
`-- COPYING
`-- Makefile
`-- <other stuff, including default.json>

This is just a basic idea; it's likely I'll change this around a lot. It is also not a bad idea to move the generated data files into the data directory to avoid adding a lot of junk to the root directory.

Other things:

• Probably could either get rid of preseed.txt or move it to some utilities directory or something similar.
• Future: where should GUI stuff (like ui file) go?
• runsim.py should be renamed to something more like ionmd.py since it will serve to be imported for the run_simulation function for the future GUI.

Add ability to have multiple cooling lasers. At the moment this is sort of possible by cleverly choosing a good value for khat, but one cannot have conterpropagating lasers this way.

Running in a virtual machine is not really a good solution. It would be nice to be able to run simulations on more computers throughout the lab which all run Windows!

Pillow

Hi,

I'm trying to run IonMD however, I was unable to get pass. I'm on Python2.7.6+GCC4.8.2+Ubuntu and installed packages required. I'm now facing 2 questions

1. By running make, it says h2xml: Command not found
2. By running python runsim.py, I got ImportError: cannot import name Params on line33.

params has been installed and updated to the latest version via pip, I'm not sure what the problem was. Could you please help me with this?

Thanks.
Fan

Using Boost.Python instead of using ctypes could help clean up the code overall (on both the Python and C++ ends) and simplify things in general.

Currently, simulation parameters are all hard coded into runsim.py. This is obviously not the best way to handle things, since it decreases code readability and increases the chance of breaking something. One possible alternative would be to use something like JSON to store simulation parameters. This way, runsim.py would conceivably never need to be touched. This would essentially require two objects:

1. One that describes all the simulation parameters that gets passed to the C++ part (i.e., everything in the Params struct.
2. Another which controls miscellaneous simulation run stuff, such as whether to loop and change parameters, what to display at the end, etc.

The JSON file would then look something like this:

{
    "control": {
        "loop": false,
        "plot_trajectory": false,
        "plot_fourier": false,
        "ionvis.display": true
    },

    "params": {
        "N": 10,
        "V": 120,
        "U": 0,
        "ccd_bins": 768,
        "ccd_extent": 1024
    }
}

Some rework will still be required since the original plan had been to pickle parameters for easy storage and loading, but this doesn't work since the Params contains pointers. There either need to be changes there or load/save functions to convert between what C++ wants and what Python wants.

There are currently a lot of problems when doing python -3 runsim.py. This should be fixed.

Python (and for that matter C++) naming convention does not follow PEP 8. This should be changed where practical.

Currently, this simply doesn't work. This needs to be investigated and fixed.

accidental duplicate