My daughter got a hand-me-down GeForce GS8400 from a friend
at school. That's like a $45 video card, yet is presumably advanced
enough to support Cuda, and therefore PyCuda. What's that?
We keep getting back to numpy arrays as a basis for computer
graphics. Linear algebra with numpy.matrix looks like a good
way to go in a digital math class.
Perry Greenfield has had much to do with getting Python
going at the Space Telescope Science Institute (STScI).
We invited him to give a lightning talk during my class,
but he was stuck in Amsterdam owing to the volcanic
ash problem in Iceland.
Numpy has a lot in common with IDL, an inhouse analysis
language for many on the Hubble project. In sharing
research data with the wider public however, it's easier if the
files might be worked on by free and open source software.
IDL ain't cheap.