[GHC] #5165: GHC doesn't optimize FP excess precision properly

[GHC]

#5165: GHC doesn't optimize FP excess precision properly
-------------------------+--------------------------------------------------
    Reporter:  ezyang    |        Owner:                             
        Type:  bug       |       Status:  new                        
    Priority:  normal    |    Milestone:                             
   Component:  Compiler  |      Version:  7.1                        
    Keywords:            |     Testcase:                             
   Blockedby:            |   Difficulty:                             
          Os:  Linux     |     Blocking:                             
Architecture:  x86       |      Failure:  Incorrect result at runtime
-------------------------+--------------------------------------------------

Comment(by simonmar):

 Correct.  Specifically:

  * on x86 without `-msse2`, the native code generator always does excess-
 precision
  * on x86_64 and x86 with `-msse2`, the NCG never does excess-precision
  * the LLVM backend probably behaves the same (though it doesn't support
 `-msse2`, as far as I know)

 I suggest we should not care about this, but we should make `-fexcess-
 precision` emit a warning that it is not doing anything.  Eventually we'll
 be using `-msse2` by default (or x86-64), and the problem will go away.

-- 
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/5165#comment:2>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler