[GHC] #14068: Loopification using join points

GHC ghc-devs at haskell.org
Fri Mar 23 17:53:42 UTC 2018


#14068: Loopification using join points
-------------------------------------+-------------------------------------
        Reporter:  nomeata           |                Owner:  nomeata
            Type:  bug               |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Compiler          |              Version:  8.0.1
      Resolution:                    |             Keywords:  JoinPoints
Operating System:  Unknown/Multiple  |         Architecture:
                                     |  Unknown/Multiple
 Type of failure:  None/Unknown      |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:  #13966 #14067     |  Differential Rev(s):  Phab:D3811
  #14827                             |
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by simonpj):

 Yes, the entire join-ceiling machinery is wonky.  I have a long-gestated
 patch about this, not yet completed alas, but this conversation may help
 me to make progress on it.

 I think float-out should not be moving join points at all -- except
 perhaps if they can go all the way to top level.  Why top level?  Well

 * Then the function it was part of before becomes smaller, and hence may
 inline.

 * At top level there is no closure to allocate.

 I suppose that if lifted to top level then it might be loopified afresh.
 Question: how beneficial is it to loopify a top-level function?  Danger:
 once loopified, it becomes non-recursive and might be inlined again --
 which would be funny but not very productive.

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14068#comment:51>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler


More information about the ghc-tickets mailing list