NaN, Inf, -Inf in floating point operations

[snim2]

Related to #35 check that all FPU operations work correctly with pathological inputs, in the translated simulator.

[snim2]

See this thread:

https://parallella.org/forums/viewtopic.php?f=23&t=908&p=5932&hilit=float+nan#p5932

and this code:

https://github.com/adapteva/epiphany-gdb/blob/epiphany-gdb-7_4/sim/epiphany/epiphany-fp.c