[GHC] #5321: Very slow constraint solving for type families

[GHC]

#5321: Very slow constraint solving for type families
----------------------------------------+-----------------------------------
    Reporter:  simonpj                  |       Owner:                                 
        Type:  bug                      |      Status:  new                            
    Priority:  low                      |   Milestone:  7.6.1                          
   Component:  Compiler (Type checker)  |     Version:  7.0.3                          
    Keywords:                           |          Os:  Unknown/Multiple               
Architecture:  Unknown/Multiple         |     Failure:  Compile-time performance bug   
  Difficulty:  Unknown                  |    Testcase:  perf/compiler/T5321Fun, T5321FD
   Blockedby:                           |    Blocking:                                 
     Related:                           |  
----------------------------------------+-----------------------------------

Comment(by dimitris):

 With the changes I've made things have been improved. With functional
 dependencies:
 {{{
 testHMapFoo1    0.21sec            96MB
 testHMapFoo2    0.21sec            95MB
 testHMapFoo3    0.20sec            92MB
 }}}
 I dont understand very well your third allocation number Simon (you meant
 Mbytes?)
 With type families we get:
 {{{
 testHMapFoo1    0.21sec           117MB
 testHMapFoo2    0.29sec           124MB
 testHMapFoo3    0.29sec           125MB
 }}}
 So they are much more comparable (and faster too!). The slightly higher
 allocation I believe is probably caused because with type families there's
 evidence around whereas with Derived constraints from fundeps that's not
 the case.

 This all is in my branch but I will push to HEAD shortly. The file is
 already in the testuite.

-- 
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/5321#comment:5>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler