#9476: Implement late lambda-lifting -------------------------------------+------------------------------------- Reporter: simonpj | Owner: sgraf Type: feature request | Status: closed Priority: normal | Milestone: 8.8.1 Component: Compiler | Version: 7.8.2 Resolution: fixed | Keywords: LateLamLift Operating System: Unknown/Multiple | Architecture: Type of failure: Runtime | Unknown/Multiple performance bug | Test Case: Blocked By: | Blocking: Related Tickets: #8763 #13286 | Differential Rev(s): Phab:D5224 Wiki Page: LateLamLift | -------------------------------------+------------------------------------- Comment (by sgraf): I already wrote something down in [https://github.com/sgraf812/late-lam- lift/blob/master/paper.pdf section 4 here]. Table 2 turns off closure growth, in particular. The results are fairly inconclusive, but I argue that we want to avoid arbitrary increases in allocations in order to stay predictable. For example `wheel-sieve1`, with a nearly 30% increase in allocation, had an increase in residency by 5.7%. I imagine that there are programs that could trigger some exponential worst case. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9476#comment:73 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler