[GHC] #9221: (super!) linear slowdown of parallel builds on 40 core machine

[GHC]

#9221: (super!) linear slowdown of parallel builds on 40 core machine
-------------------------------------+-------------------------------------
        Reporter:  carter            |                   Owner:
            Type:  bug               |                  Status:  new
        Priority:  high              |               Milestone:  7.12.1
       Component:  Compiler          |                 Version:  7.8.2
      Resolution:                    |                Keywords:
Operating System:  Unknown/Multiple  |            Architecture:
 Type of failure:  Compile-time      |  Unknown/Multiple
  performance bug                    |               Test Case:
      Blocked By:                    |                Blocking:
 Related Tickets:  #910, #8224       |  Differential Revisions:
-------------------------------------+-------------------------------------

Comment (by simonmar):

 @bgamari, correct - ThreadScope won't tell us anything about the origin of
 blackholes, but it would confirm whether there is a lot of blocking on
 blackholes.  Ideally we should see all the capabilities fully busy when
 compiling a perfectly parallel workload. Note that dependencies between
 modules reduce the amount of parallelism available in general, so for a
 good test it would be best to construct an example with plenty of
 parallelism.

 @slyfox, I don't believe perf is telling us anything sensible here.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9221#comment:36>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler