[SoC] Week 4 OpenCL Integration

[Seth Price]

1) What did I do last week

Last week I worked a bit on the resampling API, but mainly on the code  
for r.sun. I've finished it, but am waiting on figuring out what to do  
with some pj_do_proj() calls. I found numerous places for minor  
optimization in the code. I think that once my OpenCL code is  
implemented successfully, someone could back-port it to the original  
code for a modest gain.

On the topic of pj_do_proj(): It finds the true south direction of a  
raster cell, because not all cells in all projections point due south.  
I've noticed that it's can be called a good number of times per cell,  
which is quite unneeded. Hopefully we figure out a better structure  
for it to be called once per cell, which would result in a speedup of  
the entire program even without OpenCL.

2) What do I plan to do this week

I'll be working mainly on implementing the resampling code in GDAL. If  
r.sun finds a good workaround for the pj_do_proj(), I'll implement  
that code and start debugging it.

3) Do I have any problems/obstacle which interfere with my work

Nothing major. I'd like to finish work on r.sun soon. I'll have plenty  
of work in the meantime in the GDAL warper. I have some tendonitis  
troubles which I'm working on addressing, which is slowing work, but  
otherwise steady progress.

~Seth