#tracpy
Fortran core of Tracmass + Python wrapping around the outside.
To get the Fortran Tracmass code ready:
make clean- Compile tracmass code in Fortran:
make f2py
Note: if the code will not compile, a first step could be to change the flag in the makefile from -m64 to -m32 if you are set up to use 32 bit instead of 64 bit.
To learn about the module Tracpy:
- Open an iPython notebook server in the tracpy directory:
ipython notebook - A webpage will open in your browser showing the available notebooks in the current directory. Open the notebook called manual
- The cells of the notebook can be run in order by pushing "shift" and "enter" together or the whole notebook can be run by selecting Cell > Run all. The notebook demonstrates how to initialize and run a numerical drifter experiment using tracpy.
- Alternatively, a static PDF version of the manual has been saved and can be viewed (manual.pdf) but not run.
For projects, it is suggested to start a separate directory with its own initialization and run file for the simulation(s). Then tracpy can be imported as a module and the run script can be run from the project. An example project can be found at https://github.com/kthyng/gisr.git.
Some more information about running on Linux machines with taskset to control what cores a given process is run on (the syntax of these commands depends on your run.py script):
Multiple instances of the simulation can be run on different cores if you have a multi-core Linux machine. To do this, use taskset:
taskset 3 python2.7 run.py > temp_log.txt &
This command runs the run.py script for a specific project in the background (due to the &), redirects the output from the screen to the text file 'temp_log.txt', and runs the process on core 3. This command can then be used for other instances by choosing a different core after the command taskset.
If a process is already running, you can check its current core using its PID:
taskset -p [PID]
Then you can move the process to a different core using
taskset -p [CORE NUMBER] [PID]