#1246: <= operators get compiled worse than ==
----------------------+-----------------------------------------------------
Reporter: duncan | Owner:
Type: bug | Status: new
Priority: normal | Milestone:
Component: Compiler | Version: 6.6
Severity: normal | Resolution:
Keywords: | Difficulty: Unknown
Testcase: | Architecture: x86
Os: Unknown |
----------------------+-----------------------------------------------------
Comment (by simonpj):
Scary. Are you suggesting that the compiler should
* Unroll the loop one time (as you have done manually)
* Spot that, once unrolled, the (n<=0) test can be replaced by (n==0)
That looks pretty hard to achieve to me. How would it know to unroll
once? Even once unrolled, spotting that you can relace n<=0 with n==0 is
not trivial; for example it is only valid because the loop only subtracts
one.
Maybe you instead suggesting that the code generator do a better job of
(<=). That looks more feasible to me.
Simon
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/1246>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
_______________________________________________
Glasgow-haskell-bugs mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
