[GHC] #7858: Fix definitions of abs/signum for Floats/Doubles

[GHC]

#7858: Fix definitions of abs/signum for Floats/Doubles
-------------------------------------+-------------------------------------
              Reporter:  lerkok      |            Owner:
                  Type:  bug         |           Status:  patch
              Priority:  normal      |        Milestone:  7.10.1
             Component:              |          Version:  7.6.3
  libraries/base                     |         Keywords:  floating point
            Resolution:              |     Architecture:  Unknown/Multiple
      Operating System:              |       Difficulty:  Unknown
  Unknown/Multiple                   |       Blocked By:
       Type of failure:              |  Related Tickets:  #9238
  None/Unknown                       |
             Test Case:              |
              Blocking:              |
Differential Revisions:              |
-------------------------------------+-------------------------------------

Comment (by rwbarton):

 Replying to [comment:8 carter]:
 > if we had a cheap Float<->Word32, Double <-> Word64, could we define
 these wrt the bit banging approach? In fact, can't we define those already
 using CMM (wrt the unlifted types)?
 Well, on x86, ideally we would use `andps`/`andpd` for `abs`, and stay
 within the SSE register set. I don't know what is the most efficient way
 to implement `signum`.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/7858#comment:9>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler