#5715: Inliner fails to inline a function, causing 20x slowdown
---------------------------------+------------------------------------------
Reporter: bos | Owner:
Type: bug | Status: new
Priority: normal | Milestone:
Component: Compiler | Version: 7.2.1
Keywords: | Os: Unknown/Multiple
Architecture: Unknown/Multiple | Failure: Runtime performance bug
Difficulty: Unknown | Testcase:
Blockedby: | Blocking:
Related: |
---------------------------------+------------------------------------------
Changes (by simonpj):
* difficulty: => Unknown
Comment:
The behaviour of INLINE should be simple and predictable. If you write
{{{
f x y = <rhs>
{-# INLINE f #-}
}}}
then `f` will be inlined into any application `(f e1 e2)`. But not into
partial applications.
Consider the example you give:
{{{
instance Variate Int8 where
uniform = uniform1 fromIntegral
uniformR = uniformRange
{-# INLINE uniform #-}
{-# INLINE uniformR #-}
}}}
* `uniformRange will not be inlined here. (Doing so would do no good.)
* Every use of `uniformR` (at type `Int8`) should replaced by
`uniformRange` (since `uniformR` has an INLINE pragma and has no args).
* If that call to `uniformR` had two arguments, then `uniformRange`
should in turn inline.
I see no reason to write
{{{
instance Variate Int8 where
uniform = uniform1 fromIntegral
uniformR inliner sacrifice = uniformRange inliner sacrifice
{-# INLINE uniform #-}
{-# INLINE uniformR #-}
}}}
Maybe that is a bug.
If you believe that HEAD (or even 7.2) is showing odd behaviour, could you
make a standalone test case? Then I will certainly look at it.
Thanks
Simon
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/5715#comment:1>
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
