[GHC] #14914: Only turn suitable targets into a fallthrough in CmmContFlowOpt.

#14914: Only turn suitable targets into a fallthrough in CmmContFlowOpt. -------------------------------------+------------------------------------- Reporter: AndreasK | Owner: (none) Type: task | Status: new Priority: normal | Milestone: Component: Compiler | Version: (CodeGen) | Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: None/Unknown Unknown/Multiple | Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- We are inverting Branches in CmmContFlowOpt such that we get as many fall through paths as possible. However as it stands besides the conditionals themselves the only criterion used is the number of predecessors for a Block. This often (but not always) aligns with the actual possibility of turning branches into a fallthrough. We should at least look at the cost/benefit of tracking actual predecessors and see if that turns out as something worthwhile. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/14914 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#14914: Only turn suitable targets into a fallthrough in CmmContFlowOpt. -------------------------------------+------------------------------------- Reporter: AndreasK | Owner: (none) Type: task | Status: new Priority: normal | Milestone: Component: Compiler | Version: (CodeGen) | Resolution: | Keywords: CodeGen Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: None/Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Changes (by AndreasK): * keywords: => CodeGen Comment: It's also worth noting that inverting conditionals has a major effect on codelayout. So this has to be considered when changing the logic behind the inversions. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/14914#comment:1 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler
participants (1)
-
GHC