[GHC] #15642: Improve the worst case performance of weak pointers

[GHC]

#15642: Improve the worst case performance of weak pointers
-------------------------------------+-------------------------------------
        Reporter:  dfeuer            |                Owner:  (none)
            Type:  feature request   |               Status:  new
        Priority:  normal            |            Milestone:  8.8.1
       Component:  Runtime System    |              Version:  8.6.1-beta1
      Resolution:                    |             Keywords:
Operating System:  Unknown/Multiple  |         Architecture:
 Type of failure:  Runtime           |  Unknown/Multiple
  performance bug                    |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:                    |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by dfeuer):

 A more conservative approach (less cost for less benefit) would be to only
 build the hash table during the first/second/third round of fixpointing.
 I'm a bit surprised we can only spare one bit for 32-bit systems. Isn't
 the tables+code area likely to be much smaller than the address space?

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15642#comment:5>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler