vbo.buffer nor vbo.buffers aren't normally used by external code
The normal way to get the ID of the VBO is int(vbo)
See https://github.com/inducer/pyopencl/issues/196

Using type suffix

Fixing CompilerWarning: <program source>: warning: double precision constant requires cl_khr_fp64, casting to single precision
An application that wants to use double will need to include the directive before any double precision data type is declared in the kernel code.

Rename pyopencl.ipython to pyopencl.ipython_ext to convince 2to3 that 'from IPython import' is not package-local

Restore Python 3 compatibility; Apply PEP8 and PyFlakes; Remove unused imports.

Fixes import error when using Pillow, a PIL fork with many bug fixes and Python 3 compatibility.

* Numpy does element-wise operations by default.  Updated the cpu
  operation to use pure numpy.

* Eliminated the loop which is not necessary to demonstrate
  parallelism on array operations.

* Made the number of workers explicit rather than gpu chosen,
  through local_size variable passed to kernel execution.

* Increased to ~8 million data points to more clearly demonstrate
  the difference between cpu and gpu based computations.