[GHC] #9304: Floating point woes; Different behavior on 32- vs 64-bit x86

[GHC]

#9304: Floating point woes; Different behavior on 32- vs 64-bit x86
-------------------------------------+------------------------------------
        Reporter:  lerkok            |            Owner:
            Type:  bug               |           Status:  new
        Priority:  normal            |        Milestone:
       Component:  Compiler          |          Version:  7.8.3
      Resolution:                    |         Keywords:  floating point
Operating System:  Unknown/Multiple  |     Architecture:  Unknown/Multiple
 Type of failure:  None/Unknown      |       Difficulty:  Unknown
       Test Case:                    |       Blocked By:  9276
        Blocking:                    |  Related Tickets:
-------------------------------------+------------------------------------

Comment (by lerkok):

 @simonpj: I'm not sure if it's a real bug or not. There's an observed
 difference in behavior on 32-bit where "-O" produces a different result
 compared to when compiled without "-O"; and that's at least mildly
 disconcerting.

 @carter has all the context, and I think his expertise will come in handy
 when he gets around to spinning up a 32-bit VM and looks at it in some
 more detail. If he concludes this is par-for-the-course; then we can
 remove the ticket.

 I might be in the minority, but I do not think the difference in the ulp
 for `Double` on 64-bit vs 32-bit a remarkably small divergence.. I wish
 `Double` meant precisely the same thing, regardless of architecture; i.e.,
 the 64-bit IEEE double-precision number. But I also understand that this
 is a minor enough concern for most Haskell users to trickle up chain of
 importance.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9304#comment:27>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler