update TaylorGreen example, replace dtAdvecCrit method_key by dtCrit, add default filename in dt_adapt operator

Fix fortran install process + shared lib by default (much more verbose ...), review default method setup in operators, review fftw curl.

24 - test_opencl_environment (Failed), because of pep8
Examples not tested but NSDebug.py which works fine.

Add baroclinic operator and associated example for RTI benchmark (which is still not working properly) + some documentation + several addings/improvements  

Code improvements:
  - Standardize MPI communicator use (not always COMM_WORKD);
  - Moving operator requirements to continuous level;
  - Add task parallelism paradigm;
  - GPU: Multi-Scale advection reorganisation on GPU (reuse of remeshing formula);
  - GPU: Reuse buffer across different discretization of the same field.

Fix bug in HDF5/VTK output
Fix bug in field initialization
Modif in FD and diff operators

general updates concerning NS benchmark with bluff bodies (flow past sphere, etc ...) - still some problems to solve ...

Adaptative vorticity projection according a given criterion + velocity correction (to prescribe proper mean flow and ensure the desired inlet flowrate) + some updates in adaptative time step procedure

improvement of the adaptative time step operator (the choice of dt_advection criterion now relies on the use of a method_key in a dictionnary)

general updates with in particular, simulation with adaptative time step, script for flow past bluff bodies, forces computation

Improvements :\n -Automatic management of redistribute waitings in destination operators.\n -Redistribute transfers (MPI and OpenCL) could be seperated along vector components.\n -OpenCL kernels events use to synchronize GPU computations. \n -Examples update.

Handle data transfers from/to devices within the Redistribute class. Update examples. Update gpu benchmarks