Potential Optimization: Apply XOR and run-length encoding

[jackiryan]

A fellow JPL-er suggested some simple things that could help speed up the data processing and the plotting in LiveView, taken from optimizations used in the old 16-bit Commodore Amiga days...

• Apply an XOR to the frame to be displayed and the previous frame. This effectively removes pixels that stay the same and then only leaves the pixels that changed. At a high level, all of the frameview widget plotting functions should only plot the pixels that changed their values from the previous frame.
• Use run-length encoding to compress the memory being sent to and from the different objects (in particular, the GPU). This is dead-simple lossless compression that just groups similar pixels together.

[jackiryan]

Looks like this did not provide much of a speed-up on the MPSoC. Most of the time is spent in the replot function...