Fix floating point comparison

[petitnau]

After #131, the fgt.* t0, t1, t2 and fge.* t0, t1, t2 pseudo-instructions are defined, respectively, as fle.* t0, t2, t1 and flt.* t0, t2, t1.
However, they should be defined as flt.* t0, t2, t1 and fle.* t0, t2, t1, as x > y <=> y < x and x >= y iff y <= x.

[TheThirdOne]

I am a little surprised that the psuedo intructions were wrong after being corrected. I was clearly not thinking about correctness when I wrote them.