[GHC] #14880: GHC panic: updateRole

[GHC]

#14880: GHC panic: updateRole
-------------------------------------+-------------------------------------
        Reporter:  RyanGlScott       |                Owner:  goldfire
            Type:  bug               |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Compiler (Type    |              Version:  8.2.2
  checker)                           |
      Resolution:                    |             Keywords:  TypeInType
Operating System:  Unknown/Multiple  |         Architecture:
 Type of failure:  Compile-time      |  Unknown/Multiple
  crash or panic                     |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:  #15076            |  Differential Rev(s):  Phab:D4769
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by tdammers):

 I did some benchmarking of `Data.Map` and also `Data.IntMap` in isolation.

 For both, I implemented combining a large number of positive integers into
 a set and then testing for the presence of `-1` in the set, using two
 strategies:

 1. the "unions" approach: `unions . map singleton`
 2. the "insert" approach: `foldr insert empty`

 For `IntMap` (which is what we're using here), the "unions" approach is
 slightly faster (~14.5s vs. ~15s for 100 million elements); for `Map`, the
 "unions" approach is a lot faster (~6.5s vs. ~8.2s for 10 million
 elements).

 Note that this is the worst possible case for `union`, and it still
 outperforms individual insertions for some reason.

 So if that is accurate, then the idea that speeding things up by using
 lots of inserts instead of lots of set unions doesn't hold up.

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14880#comment:56>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler