#4941: SpecConstr generates functions that do not use their arguments
---------------------------------+------------------------------------------
Reporter: simonpj | Owner:
Type: task | Status: new
Priority: normal | Milestone: _|_
Component: Compiler | Version: 7.0.1
Keywords: | Testcase:
Blockedby: | Difficulty:
Os: Unknown/Multiple | Blocking:
Architecture: Unknown/Multiple | Failure: Runtime performance bug
---------------------------------+------------------------------------------
Comment(by batterseapower):
Replying to [comment:17 simonpj]:
> FYI: compiling with `-ticky` is a fantastic way to nail exactly where
allocation increases.
I have just tried this without success. Ticky tells me that a function
$wtrim is increasing allocs by 23%, but I have traced it through manually
and I just can't see how that can be happening.
(I infer that the -ticky output includes those allocations made within a
let-no-escape in the allocation count of their caller).
To me it looks like the optimised code is never worse and sometimes an
improvement in # of allocations metric - and optimisation reduces the
number of let-no-escapes to boot. So I'm really flummoxed as to why
allocations increase. It doesn't seem to be just because closure sizes
increase or something because ALLOC_HEAP_ctr rises. Perhaps it would be
clearer if ticky either told me a breakdown of allocation by the type of
thing being allocated or by the let binding from which it originated.
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/4941#comment:18>
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
