[GHC] #10678: integer-gmp's runS seems unnecessarily expensive

[GHC]

#10678: integer-gmp's runS seems unnecessarily expensive
-------------------------------------+-------------------------------------
        Reporter:  rwbarton          |                   Owner:
            Type:  bug               |                  Status:  new
        Priority:  normal            |               Milestone:
       Component:  Compiler          |                 Version:  7.10.1
  (CodeGen)                          |
      Resolution:                    |                Keywords:
Operating System:  Unknown/Multiple  |            Architecture:
 Type of failure:  Runtime           |  Unknown/Multiple
  performance bug                    |               Test Case:
      Blocked By:                    |                Blocking:
 Related Tickets:                    |  Differential Revisions:
-------------------------------------+-------------------------------------

Comment (by simonpj):

 Interesting.  DO do this less manually we would need two things:

  * The expression `runRW e` has the CPR property if `e rw` has the nested
 CPR property.

  * That might lead us to w/w a function like `plusBigNat` which in turn
 would lead to expressions like
 {{{
    case (runRW e) of BN# farr -> farr
 }}}
   we would need some kind of special case to push the `case` inside the
 `runRW` to get
 {{{
   runRW (\s -> case e s of (# s', r #) ->
                case r of BN# farr -> (# s', farr #)
 }}}
   That does not look too hard.

 It'd be good to resurrect nested CPR (there's a ticket for that, #1600).

 Simon

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10678#comment:10>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler