[GHC] #14667: Compiling a function with a lot of alternatives bottlenecks on insertIntHeap

[GHC]

#14667: Compiling a function with a lot of alternatives bottlenecks on
insertIntHeap
-------------------------------------+-------------------------------------
        Reporter:  niteria           |                Owner:  niteria
            Type:  bug               |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Compiler          |              Version:
      Resolution:                    |             Keywords:
Operating System:  Unknown/Multiple  |         Architecture:
 Type of failure:  Compile-time      |  Unknown/Multiple
  performance bug                    |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:                    |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by simonmar):

 That diff of mine basically replaces a set with a list, so it's not
 entirely surprising that it can be responsible for something being
 accidentally quadratic.  I think at the time it was a small improvement
 because N was small in everything I measured, and for small N a list is a
 bit faster than an IntSet, but again it's not entirely surprising to find
 that it's possible to make it behave badly.

 I'm totally fine with reverting it, but do a nofib run first to check for
 adverse effects.

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14667#comment:6>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler