[GHC] #693: dynamic locking

[GHC]

#693: dynamic locking
-------------------------------+--------------------------------------------
  Reporter:  simonmar          |          Owner:         
      Type:  task              |         Status:  new    
  Priority:  normal            |      Milestone:  6.6.1  
 Component:  Runtime System    |        Version:  6.4.1  
  Severity:  normal            |     Resolution:         
  Keywords:                    |             Os:  Unknown
Difficulty:  Moderate (1 day)  |   Architecture:  Unknown
-------------------------------+--------------------------------------------
Changes (by simonmar):

  * milestone:  6.6 => 6.6.1

Old description:

> Now that the SMP and threaded runtimes are the same, there is some
> unnecessary locking going on in the single threaded case.  For example,
> MVar operations now require an atomic memory operation, which can be
> quite slow.  It might be a good idea to dynamically turn on these locks
> if we're running multi-threaded (ie. +RTS -N2 or greater).  ToDo: measure
> the difference.

New description:

 Now that the SMP and threaded runtimes are the same, there is some
 unnecessary locking going on in the single threaded case.  For example,
 MVar operations now require an atomic memory operation, which can be quite
 slow.  It might be a good idea to dynamically turn on these locks if we're
 running multi-threaded (ie. +RTS -N2 or greater).  !ToDo: measure the
 difference.

-- 
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/693>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler