From the webpage:
Galry is a high performance interactive 2D visualization package in Python. It lets you visualize and navigate into very large 2D plots (signals, points, textures…) in real time, by using the graphics card as much as possible (with OpenGL). On a 2012 computer with a recent 250$ graphics card, one can interactively visualize as much as 100 million points at a reasonable framerate.
Galry is not meant to generate high-quality plots (like matplotlib), and is more “low-level”. It can be used to write complex interactive visualization GUIs that deal with large 2D datasets (only with QT for now).
It is based on PyOpenGL and Numpy and is meant to work on any platform (Window/Linux/MacOS). Mandatory dependencies include Python 2.7, Numpy, either PyQt4 or PySide, PyOpenGL, matplotlib.
Optional dependencies include IPython, hdf5, PyOpenCL (the last two are not currently used but may be in the future).
Important note: Galry is still an experimental project with an unstable programming interface that is likely to change at any time. Do not use it in production yet.
If you need 2D visualization performance, take a look at Galry.
The demo video is impressive to say the least.