[GHC] #5113: Huge performance regression of 7.0.2, 7.0.3 and HEAD over 7.0.1 and 6.12 (MonoLocalBinds)

[GHC]

#5113: Huge performance regression of 7.0.2, 7.0.3 and HEAD over 7.0.1 and 6.12
(MonoLocalBinds)
-------------------------------------+-------------------------------------
        Reporter:                    |                   Owner:
  daniel.is.fischer                  |                  Status:  closed
            Type:  bug               |               Milestone:
        Priority:  normal            |                 Version:  7.0.3
       Component:  Compiler          |                Keywords:
      Resolution:  fixed             |  performance, MonoLocalBinds
Operating System:  Linux             |            Architecture:  x86
 Type of failure:  Runtime           |               Test Case:
  performance bug                    |  perf/should_run/T5113
      Blocked By:                    |                Blocking:
 Related Tickets:                    |  Differential Revisions:
-------------------------------------+-------------------------------------

Comment (by bgamari):

 It seems to me like the fix for #10527 is a bit questionable. I may be
 missing something but I don't see how arguments could possibly get to the
 rebuilder (which is responsible for trying rewrite rules) as we now simply
 perform a bunch of substitutions on them with `substExprS` instead of
 passing them to`simplExpr`.

 Surely I'm missing something as one would think this would result in
 massive a performance regression if it really were the case. Simon,
 perhaps you can explain how you intended this to work?

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/5113#comment:11>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler