[GHC] #13395: 3x slowdown on GHC HEAD with file containing lots of overloaded string literals

[GHC]

#13395: 3x slowdown on GHC HEAD with file containing lots of overloaded string
literals
-------------------------------------+-------------------------------------
        Reporter:  RyanGlScott       |                Owner:  (none)
            Type:  bug               |               Status:  closed
        Priority:  high              |            Milestone:  8.2.1
       Component:  Compiler          |              Version:  8.1
      Resolution:  invalid           |             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:                    |
-------------------------------------+-------------------------------------
Changes (by RyanGlScott):

 * status:  new => closed
 * resolution:   => invalid


Comment:

 Ugh, it turns out this was caused by me using a GHC HEAD that was built
 with the `quick` flavor instead of the `perf` flavor. I had thought that
 the only difference between the two was that the latter built Haddocks and
 other non-essentials—that couldn't have been more wrong!

 After rebuilding GHC HEAD with `perf`, I get this time:

 {{{
 $ /usr/bin/time ~/Software/ghc4/inplace/bin/ghc-stage2 -fforce-recomp
 TextXmlHtmlHTMLMeta.hs
 [1 of 1] Compiling Text.XmlHtml.HTML.Meta ( TextXmlHtmlHTMLMeta.hs,
 TextXmlHtmlHTMLMeta.o )
 3.08user 0.07system 0:03.17elapsed 99%CPU (0avgtext+0avgdata
 196820maxresident)k
 0inputs+21464outputs (0major+41873minor)pagefaults 0swaps
 }}}

 Which is even //better// than 8.0.2! So this tale had a happy ending after
 all.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13395#comment:3>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler