#2915: Arity is smaller than need be --------------------------------------------+------------------------------ Reporter: simonpj | Owner: simonpj Type: bug | Status: closed Priority: lowest | Milestone: 7.6.2 Component: Compiler | Version: 6.10.1 Resolution: fixed | Keywords: Operating System: Unknown/Multiple | Architecture: Type of failure: Runtime performance bug | Unknown/Multiple Test Case: | Difficulty: Unknown Blocking: | Blocked By: | Related Tickets: --------------------------------------------+------------------------------ Changes (by nomeata): * status: infoneeded => closed * resolution: => fixed Comment: It seems that not test cases fail with this change, and only one nofib benchmark profits (mandel -1.9%). It would also simplify `CoreArity` (no need for the `ae_bndrs` field). (Or course the scrunitee must not be expensive; I guard this using the `ae_cheap_fn`. If I don't do that, fluid improves allocation by -2.3%, but of course it's wrong. In most cases, though, FloatOut will have ensured that the scrutinee is cheap) I’ll need to run another round of nofib, with an unloaded machine, to ensure that the runtimes are still ok... and here it is {{{ Min -0.1% -1.9% -3.0% -2.9% 0.0% Max +0.0% +0.0% +3.5% +2.2% +3.8% Geometric Mean -0.0% -0.0% -0.4% -0.4% +0.0% }}} Runtime varies a bit, with a slight tendency to an improvement. Guess that is good enough to push. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/2915#comment:13 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler