[GHC] #14827: Recognize when inlining would create a join point

[GHC]

#14827: Recognize when inlining would create a join point
-------------------------------------+-------------------------------------
        Reporter:  ersetzen          |                Owner:  (none)
            Type:  feature request   |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Compiler          |              Version:  8.2.2
      Resolution:                    |             Keywords:  JoinPoints
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:                    |
-------------------------------------+-------------------------------------

Comment (by simonpj):

 I'm too swamped to follow up why `f` doesn't get inlined on the branch,
 but I know that Joachim is planning to get back to loopification.

 >  Side note: Neither the loopification branch and ghc head run faster
 with ghci.

 The Description says that ghci is 20x faster (in 8.2.2).  So in HEAD do we
 have the 20x fast performance in both cases (ie good), or do we have the
 1x performance in both cases (ie bad)?

 I still don't understand why GHCi wins so big in any version.

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14827#comment:8>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler