[GHC] #13739: Very slow linking of profiled executables

[GHC]

#13739: Very slow linking of profiled executables
-------------------------------------+-------------------------------------
        Reporter:  duog              |                Owner:  (none)
            Type:  bug               |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Compiler          |              Version:  8.2.1-rc2
      Resolution:                    |             Keywords:
Operating System:  Linux             |         Architecture:  x86_64
 Type of failure:  Compile-time      |  (amd64)
  performance bug                    |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:                    |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------
Changes (by RyanGlScott):

 * cc: rwbarton (added)


Comment:

 Commit 283acec1d7307fdbd8cd7b3f1d984a036366d6b4 (Make split sections by
 default work again) is what caused the other regression. Before that
 commit, I obtained these times with duog's script:

 {{{
 (Without -fwhole-archive-hs-libs)
 real    0m0.306s
 user    0m0.260s
 sys     0m0.028s

 (With -fwhole-archive-hs-libs)
 real    0m0.056s
 user    0m0.036s
 sys     0m0.012s
 }}}

 Notice that the `-fwhole-archive-hs-libs` times are pretty much identical
 to the GHC 8.0.2 times I found in comment:9.

 After that commit, I obtained:

 {{{
 (Without -fwhole-archive-hs-libs)
 real    0m4.656s
 user    0m4.548s
 sys     0m0.088s

 (With -fwhole-archive-hs-libs)
 real    0m0.056s
 user    0m0.040s
 sys     0m0.008s
 }}}

 Which is basically identical to the times I found pre-
 b207b536ded40156f9adb168565ca78e1eef2c74.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13739#comment:10>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler