[GHC] #8766: length [Integer] is twice as slow but length [Int] is 10 times faster

[GHC]

#8766: length [Integer] is twice as slow but length [Int] is 10 times faster
--------------------------------------------+------------------------------
        Reporter:  George                   |            Owner:
            Type:  bug                      |           Status:  new
        Priority:  normal                   |        Milestone:
       Component:  Compiler                 |          Version:  7.8.1-rc1
      Resolution:                           |         Keywords:
Operating System:  Unknown/Multiple         |     Architecture:
 Type of failure:  Runtime performance bug  |  Unknown/Multiple
       Test Case:                           |       Difficulty:  Unknown
        Blocking:                           |       Blocked By:
                                            |  Related Tickets:
--------------------------------------------+------------------------------

Comment (by hvr):

 Replying to [comment:2 nomeata]:
 > The improvement for `Int` is most likely from 82f56e5 (#876). But I
 guess the important point of the ticket is the regression for `Integer`.

 I was getting to that (needed to verify something first though):

 The Core for

 {{{#!hs
 intlen = length [1..(2^(30::Int))::Integer]
 }}}


 Otoh, looks differs as following:

 GHC 7.6.3:
 {{{
 intlen1 = enumDeltaToInteger intlen3 intlen3 intlen2
 }}}

 vs.

 GHC 7.8.20140130:
 {{{
 intlen1 = enumDeltaToIntegerFB (incLen) I# intlen3 intlen3 intlen2
 }}}

 The use of `enumDeltaToIntegerFB` instead of `enumDeltaToInteger` accounts
 for the speed difference. I've verified that by copying the definition
 from `GHC.Enum` into the test-code and using `enumDeltaToInteger` directly
 with GHC 7.8.20140130 which resulted in a slightly better runtime than for
 GHC 7.6.3

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8766#comment:4>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler