#2422: Unrelated instance foils optimizer
--------------------------+-------------------------------------------------
Reporter: sedillard | Owner:
Type: bug | Status: new
Priority: normal | Component: Compiler
Version: 6.8.3 | Severity: normal
Keywords: | Testcase:
Architecture: x86 | Os: Linux
--------------------------+-------------------------------------------------
I have a type `(:.) a b` that I use for vectors, so a 4d vector is
`a:.a:.a:.a:.()`. This is just a linked list, so I also have a packed
representation for Doubles, `Vec4D` and a class that relates the two
representations.
The Storable instances for `(:.)` are inductive: if `a` and `(b:.c)` are
Storable then so is `(a:.b:.c)`. So long as the source is a peek and the
destination is a poke, the method calls optimize neatly and no `(:.)`
values are allocated.
Here's the weird part: If I use the Storable methods of `(:.)` and the
pack/unpack methods to extend a `Storable` instance to `Vec4D`, then the
optimization of ''the `(:.)` Storable methods'' fails. The `Vec4D`
instances are not used at all, and yet they affect optimization of the
`(:.)` instances.
This is probably related to #2416. If I move the `Storable` instance for
`(:.)` into `Main.hs`, then all optimization fails regardless of any
instance for `Vec4D`. If I move the `Storable` instance for `Vec4D` into a
different module than the instance for `(:.)`, then optimization works
fine.
The example is in two files. I tried to merge them into one but that's
when I discovered #2416.
Compiled with -O2.
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/2422>
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
