[GHC] #10397: Compiler performance regression 7.6 -> 7.8 in elimCommonBlocks

[GHC]

#10397: Compiler performance regression 7.6 -> 7.8 in elimCommonBlocks
-------------------------------------+-------------------------------------
        Reporter:  TobyGoodwin       |                   Owner:
            Type:  bug               |                  Status:  new
        Priority:  normal            |               Milestone:
       Component:  Compiler          |                 Version:  7.8.4
      Resolution:                    |                Keywords:
Operating System:  Unknown/Multiple  |  performance
 Type of failure:  None/Unknown      |            Architecture:
      Blocked By:                    |  Unknown/Multiple
 Related Tickets:                    |               Test Case:  see ticket
                                     |                Blocking:
                                     |  Differential Revisions:
-------------------------------------+-------------------------------------

Comment (by nomeata):

 Compiling your `RegBig.hs`. Without `elimCommonBlocks`:

 > <<ghc: 30310299032 bytes, 1461 GCs, 80330740/286438424 avg/max bytes
 residency (30 samples), 786M in use, 0.001 INIT (0.000 elapsed), 20.887
 MUT (33.307 elapsed), 10.015 GC (10.013 elapsed) :ghc>>

 Before:

 > <<ghc: 97671621696 bytes, 2579 GCs, 77685892/300385928 avg/max bytes
 residency (31 samples), 777M in use, 0.001 INIT (0.001 elapsed), 134.310
 MUT (146.546 elapsed), 10.639 GC (10.637 elapsed) :ghc>>

 after my patch:

 > <<ghc: 69883688328 bytes, 3038 GCs, 83489944/273420096 avg/max bytes
 residency (39 samples), 750M in use, 0.001 INIT (0.001 elapsed), 55.414
 MUT (67.757 elapsed), 14.896 GC (14.892 elapsed) :ghc>>

 So it still takes long, but the time spend in `elimCommonBlocks` is cut
 down by ⅔.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10397#comment:9>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler