[GHC] #14610: newtype wrapping of a monadic stack kills performance

[GHC]

#14610: newtype wrapping of a monadic stack kills performance
-------------------------------------+-------------------------------------
        Reporter:  mrkkrp            |                Owner:  (none)
            Type:  bug               |               Status:  new
        Priority:  normal            |            Milestone:  8.6.1
       Component:  Compiler          |              Version:  8.2.2
      Resolution:                    |             Keywords:
Operating System:  Unknown/Multiple  |         Architecture:
 Type of failure:  Runtime           |  Unknown/Multiple
  performance bug                    |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:                    |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------
Changes (by nomeata):

 * cc: nomeata (added)


Comment:

 The polymorphic `go` in comment:3 is a separate problem from the one that
 I described further up (but it might well be the OP’s problem). It could
 be fixed by static argument transformation. On my TODO list, after #14068,
 is SAT for join points, but it would not help here, because only after SAT
 it is a join point. Exitification unfortunately also does not help to turn
 this into a joint point. Maybe worth opening a separate ticket.

 Your example in comment:4 is precisely what I described, thanks for
 working it out. The fix here probably requires some significant thought
 about the join point typing. Maybe it is enough to relax the rule “The
 return type must not depend on any arguments” to “The return type must be
 representatoinally equal for all arguments” and that might allow us to
 {{{e `cast` co}}} to be a tail-call position.

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14610#comment:6>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler