#1969: enormous compile times
------------------------------------------+---------------------------------
Reporter: duncan | Owner:
Type: compile-time performance bug | Status: new
Priority: normal | Milestone: 6.8.3
Component: Compiler | Version: 6.8.1
Severity: normal | Resolution:
Keywords: performance | Difficulty: Difficult (1 week)
Testcase: | Architecture: Multiple
Os: Multiple |
------------------------------------------+---------------------------------
Comment (by igloo):
This program:
{{{
main :: IO ()
main = writeFile "W.hs" $ unlines $ map unlines foo
foo :: [[String]]
foo = [ "module J where",
"class C a where",
" c :: a -> String",
" d :: a -> String",
" d x = c x",
" e :: a -> String",
" e x = c x"
] :
[ ["data " ++ d ++ " = " ++ d,
"instance C " ++ d ++ " where",
" c " ++ d ++ " = \"" ++ d ++ "\""]
| i <- [1..1000],
let d = 'A' : show i
]
}}}
generates a `W.hs` with lots of instances like
{{{
data A24 = A24
instance C A24 where
c A24 = "A24"
}}}
Compiling with
{{{
ghc -fforce-recomp -O -c W.hs +RTS -p
}}}
this is the interesting bit of the profile:
{{{
individual
COST CENTRE MODULE no. entries %time %alloc
specProgram Specialise 8794 2006 0.0 0.0
specBind Specialise 8804 6006 0.0 0.0
splitUDs Specialise 8883 2002 3.1 4.0
elemVarSet VarSet 9174 1998000 0.1 0.2
listToCallDetails Specialise 8985 2000002 1.0 2.3
unionCalls Specialise 9182 1998000 31.1 37.6
tcCmpType Type 9186 57864047 9.3 0.0
intersectsVarSet VarSet 8885 4006001 0.2 0.0
}}}
The number of entries for the last 5 grows quadratically with the number
of instances.
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/1969#comment:3>
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
