[GHC] #10482: Not enough unboxing happens on data-family function argument

GHC ghc-devs at haskell.org
Mon Jun 29 08:56:10 UTC 2015


#10482: Not enough unboxing happens on data-family function argument
-------------------------------------+-------------------------------------
        Reporter:  akio              |                   Owner:
            Type:  bug               |                  Status:  new
        Priority:  normal            |               Milestone:
       Component:  Compiler          |                 Version:  7.10.1
      Resolution:                    |                Keywords:
Operating System:  Unknown/Multiple  |            Architecture:
 Type of failure:  Runtime           |  Unknown/Multiple
  performance bug                    |               Test Case:
      Blocked By:                    |                Blocking:
 Related Tickets:                    |  Differential Revisions:
-------------------------------------+-------------------------------------

Comment (by simonpj):

 The patch in comment:2 fixes the original bug report but gives perf
 failures on
 {{{
 >       haddock.Cabal
 >       haddock.compiler
 }}}
 Joachim says: I can confirm this, if you look at
 https://perf.haskell.org/ghc/ you see that that commit is the only in red

 Here is the report for that commit:
 https://perf.haskell.org/ghc/#revision/0b7e538a09bc958474ec704063eaa08836e9270e
 and you can fetch the numbers from there, or from the full build log at
 https://raw.githubusercontent.com/nomeata/ghc-speed-
 logs/master/0b7e538a09bc958474ec704063eaa08836e9270e.log

 It looks very reproducible as well:
 https://perf.haskell.org/ghc/#graph/tests/alloc/haddock.Cabal;hl=0b7e538a09bc958474ec704063eaa08836e9270e

 So we have to discover why this is happening. Sigh.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10482#comment:4>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler


More information about the ghc-tickets mailing list