So we don't know how much of the speed increase is due to fixing the bug, and how much is due to inlining the array alloc. Still, it's
#8885: Add inline versions of clone array primops -------------------------------------+------------------------------------ Reporter: tibbe | Owner: simonmar Type: feature request | Status: patch Priority: normal | Milestone: Component: Compiler | Version: 7.9 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Unknown/Multiple Type of failure: None/Unknown | Difficulty: Unknown Test Case: | Blocked By: Blocking: | Related Tickets: -------------------------------------+------------------------------------ Comment (by tibbe): Replying to [comment:5 simonmar]: probably better to inline the array alloc. Right. Is the bug in the implementation of `allocate` or in the implementation of the primops? If it's in the primops this bug exists in at least `stg_newArrayzh` as well, as I stole the allocation code from there. If you let me know when the bug is fixed, I will rerun the benchmark. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/8885#comment:6 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler