Accelerated version of the friendly PIL fork by Alex Clark and Contributors. For SSE4 capable x86-64 processors.