We have developed specialized routines to make use of NVIDIA CUDA GPU processing to speed up some operations (e.g. FIR filtering) by up to 10x. If you want to use NVIDIA CUDA, you should install:
For example, on Ubuntu 15.10, a combination of system packages and git
packages can be used to install the CUDA stack:
# install system packages for CUDA
$ sudo apt-get install nvidia-cuda-dev nvidia-modprobe
# install PyCUDA
$ git clone http://git.tiker.net/trees/pycuda.git
$ cd pycuda
$ ./configure.py --cuda-enable-gl
$ git submodule update --init
$ make -j 4
$ python setup.py install
# install skcuda
$ cd ..
$ git clone https://github.com/lebedov/scikit-cuda.git
$ cd scikit-cuda
$ python setup.py install
To initialize mne-python cuda support, after installing these dependencies and running their associated unit tests (to ensure your installation is correct) you can run:
$ MNE_USE_CUDA=true MNE_LOGGING_LEVEL=info python -c "import mne; mne.cuda.init_cuda()"
Enabling CUDA with 1.55 GB available memory
If you have everything installed correctly, you should see an INFO-level log message telling you your CUDA hardware’s available memory. To have CUDA initialized on startup, you can do:
>>> mne.utils.set_config('MNE_USE_CUDA', 'true')
You can test if MNE CUDA support is working by running the associated test:
$ nosetests mne/tests/test_filter.py
If MNE_USE_CUDA=true and all tests pass with none skipped, then
MNE-Python CUDA support works.
In IPython, we strongly recommend using the Qt matplotlib backend for fast and correct rendering:
$ ipython --matplotlib=qt
On Linux, for example, QT is the only matplotlib backend for which 3D rendering will work correctly. On Mac OS X for other backends certain matplotlib functions might not work as expected.
To take full advantage of MNE-Python’s visualization capacities in combination with IPython notebooks and inline displaying, please explicitly add the following magic method invocation to your notebook or configure your notebook runtime accordingly:
In [1]: %matplotlib inline
If you use another Python setup and you encounter some difficulties please report them on the MNE mailing list or on github to get assistance.
If you run into trouble when visualizing source estimates (or anything else
using mayavi), you can try setting the ETS_TOOLKIT environment variable:
>>> import os
>>> os.environ['ETS_TOOLKIT'] = 'qt4'
>>> os.environ['QT_API'] = 'pyqt'
This will tell Traits that we will use Qt with PyQt bindings.
If you get an error saying:
ValueError: API 'QDate' has already been set to version 1
you have run into a conflict with Traits. You can work around this by telling the interpreter to use QtGui and QtCore from pyface:
>>> from pyface.qt import QtGui, QtCore
This line should be added before any imports from mne-python.
For more information, see http://docs.enthought.com/mayavi/mayavi/building_applications.html.