[GHC] #8345: A more efficient atomicModifyIORef'

[GHC]

#8345: A more efficient atomicModifyIORef'
-------------------------------------+-------------------------------------
              Reporter:  parcs       |            Owner:
                  Type:  bug         |           Status:  new
              Priority:  normal      |        Milestone:  7.10.1
             Component:              |          Version:  7.6.3
  libraries/base                     |         Keywords:  IORef
            Resolution:              |     Architecture:  Unknown/Multiple
      Operating System:              |       Difficulty:  Moderate (less
  Unknown/Multiple                   |  than a day)
       Type of failure:  Runtime     |       Blocked By:
  performance bug                    |  Related Tickets:  #5926
             Test Case:              |
              Blocking:              |
Differential Revisions:              |
-------------------------------------+-------------------------------------

Comment (by dfeuer):

 Replying to [comment:14 simonmar]:
 > @hvr's version is basically just a more concise way to write @parcs'
 proposed version.  I'm happy with either.  Someone want to put up a diff?

 I'll be happy to do that. I realized today that if I replace `evaluate`
 with its specification, I do indeed get the same Core for parcs's and
 hvr's versions, which of course (conditionally) proves that what you say
 is true. But using `evaluate` itself gives somewhat different core,
 because it uses `seq#` to order things properly instead of `case`. I
 haven't inspected the Cmm or asm output, so I don't know if this makes any
 difference.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8345#comment:15>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler