March 2014 - ghc-tickets

Re: [GHC] #5925: Add inline version of newArray#
by GHC 09 Mar '14

09 Mar '14

#5925: Add inline version of newArray# -------------------------------------+------------------------------------ Reporter: tibbe | Owner: simonmar Type: feature request | Status: patch Priority: normal | Milestone: 7.6.2 Component: Compiler | Version: 7.4.1 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Unknown/Multiple Type of failure: None/Unknown | Difficulty: Unknown Test Case: | Blocked By: 4258 Blocking: | Related Tickets: -------------------------------------+------------------------------------ Comment (by tibbe): Also added a version for byte arrays. See second patch. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/5925#comment:20> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

Re: [GHC] #5925: Add inline version of newArray#
by GHC 09 Mar '14

09 Mar '14

#5925: Add inline version of newArray# -------------------------------------+------------------------------------ Reporter: tibbe | Owner: simonmar Type: feature request | Status: patch Priority: normal | Milestone: 7.6.2 Component: Compiler | Version: 7.4.1 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Unknown/Multiple Type of failure: None/Unknown | Difficulty: Unknown Test Case: | Blocked By: 4258 Blocking: | Related Tickets: -------------------------------------+------------------------------------ Comment (by tibbe): I managed to replicate a bug (#8867) in the out-of-line implementation. I've updated the attachment so the inline version doesn't have this bug. Now the Cmm generated for the example Haskell code above is: {{{ Main.newArray1_entry() // [] { [(c3fD, Main.newArray1_info: const 4294967299; const 0; const 15;)] } {offset c3fD: Hp = Hp + 160; if (Hp > I64[BaseReg + 856]) goto c3fH; else goto c3fG; c3fH: I64[BaseReg + 904] = 160; R1 = PicBaseReg + Main.newArray1_closure; call (I64[BaseReg - 8])(R1) args: 8, res: 0, upd: 8; c3fG: I64[Hp - 152] = I64[PicBaseReg + stg_MUT_ARR_PTRS_DIRTY_info@GOTPCREL]; I64[Hp - 144] = 16; I64[Hp - 136] = 17; _c3fr::I64 = Hp - 152; _c3fs::I64 = _c3fr::I64 + 24; goto c3ft; c3ft: if (_c3fs::I64 < (_c3fr::I64 + 152)) goto c3fv; else goto c3fu; c3fv: I64[_c3fs::I64] = PicBaseReg + (GHC.Tuple.()_closure+1); _c3fs::I64 = _c3fs::I64 + 8; goto c3ft; c3fu: call MO_Memset(_c3fs::I64, 0, 1, 8); R1 = PicBaseReg + (GHC.Tuple.()_closure+1); call (P64[Sp])(R1) args: 8, res: 0, upd: 8; } } }}} One question: even if the card table only uses one byte, do I have to zero out the whole word allocated for it? -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/5925#comment:19> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

Re: [GHC] #5925: Add inline version of newArray#
by GHC 09 Mar '14

09 Mar '14

#5925: Add inline version of newArray# -------------------------------------+------------------------------------ Reporter: tibbe | Owner: simonmar Type: feature request | Status: patch Priority: normal | Milestone: 7.6.2 Component: Compiler | Version: 7.4.1 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Unknown/Multiple Type of failure: None/Unknown | Difficulty: Unknown Test Case: | Blocked By: 4258 Blocking: | Related Tickets: -------------------------------------+------------------------------------ Comment (by tibbe): I've updated the patch to try to split out the array closure allocation code into a helper function. I'm currently trying to get validate to pass. If it helps reviewing, here's an output example. Input function: {{{#!haskell newArray :: IO () newArray = IO $ \s -> case newArray# 16# () s of (# s', arr #) -> (# s', () #) }}} Output Cmm: {{{ Main.newArray1_entry() // [] { [(c3fD, Main.newArray1_info: const 4294967299; const 0; const 15;)] } {offset c3fD: Hp = Hp + 160; if (Hp > I64[BaseReg + 856]) goto c3fH; else goto c3fG; c3fH: I64[BaseReg + 904] = 160; R1 = PicBaseReg + Main.newArray1_closure; call (I64[BaseReg - 8])(R1) args: 8, res: 0, upd: 8; c3fG: I64[Hp - 152] = I64[PicBaseReg + stg_MUT_ARR_PTRS_DIRTY_info@GOTPCREL]; I64[Hp - 144] = 16; I64[Hp - 136] = 17; _c3fr::I64 = Hp - 152; _c3fs::I64 = _c3fr::I64 + 24; goto c3ft; c3ft: if (_c3fs::I64 < (_c3fr::I64 + 128)) goto c3fv; else goto c3fu; c3fv: I64[_c3fs::I64] = PicBaseReg + (GHC.Tuple.()_closure+1); _c3fs::I64 = _c3fs::I64 + 8; goto c3ft; c3fu: call MO_Memset(_c3fs::I64, 0, 1, 8); R1 = PicBaseReg + (GHC.Tuple.()_closure+1); call (P64[Sp])(R1) args: 8, res: 0, upd: 8; } } }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/5925#comment:18> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

Re: [GHC] #7478: setSessionDynFlags does not always work
by GHC 08 Mar '14

08 Mar '14

#7478: setSessionDynFlags does not always work -------------------------------------+------------------------------------ Reporter: edsko | Owner: Type: bug | Status: new Priority: highest | Milestone: 7.8.1 Component: Compiler | Version: 7.7 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Unknown/Multiple Type of failure: None/Unknown | Difficulty: Unknown Test Case: ghc-api/T7478 | Blocked By: Blocking: | Related Tickets: -------------------------------------+------------------------------------ Comment (by bennofs): Replying to [comment:11 MikolajKonarski]: > BTW, right now (I don't know how it was before) calling `setSessionDynFlags` causes significant memory leaks, visible already in tests with a couple hundred modules and calls. If anybody is interested, I can try to extract a test from my code that uses GHC API, but it's GHC version dependent, etc., so I'd rather not do it until anybody is actively looking at this. But it's straightforward: basically, in a loop: > > {{{ > setSessionDynFlags flags > setTargets yetAnotherSIngleModule > load LoadAllTargets > }}} > > > > Even if we load 100 of modules and then just keep reloading the same module, memory will increase all the time. I can confirm this. I have a workaround for that problem in https://github.com/bennofs/ghc- server/blob/master/src/Server/Configure.hs#L117, which suggests that the problem is with reloading the package database. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/7478#comment:12> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

[GHC] #8863: ghc 7.6.3: type parser accepts => as -> (sometimes)
by GHC 08 Mar '14

08 Mar '14

#8863: ghc 7.6.3: type parser accepts => as -> (sometimes) --------------------------+------------------------------------------------ Reporter: | Owner: patrikj | Status: new Type: bug | Milestone: Priority: normal | Version: 7.6.3 Component: | Operating System: Linux Compiler | Type of failure: GHC accepts invalid program Keywords: | Test Case: Architecture: x86_64 | Blocking: (amd64) | Difficulty: | Unknown | Blocked By: | Related Tickets: | --------------------------+------------------------------------------------ The following examples are all parsed and type checked even though they don't seem to be conforming to the Haskell standard syntax: a :: Int => Int a = (1+) b :: a -> a => a b = const c :: Num a => [a => a => a] c = [(+)] Slight variations fail (as the should) - for example b' :: a -> b => a b' = const /Patrik ---- Linux cse-814009 3.11.0-18-generic #32-Ubuntu SMP Tue Feb 18 21:11:14 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8863> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 1

Re: [GHC] #5925: Add inline version of newArray#
by GHC 08 Mar '14

08 Mar '14

#5925: Add inline version of newArray# -------------------------------------+------------------------------------ Reporter: tibbe | Owner: simonmar Type: feature request | Status: patch Priority: normal | Milestone: 7.6.2 Component: Compiler | Version: 7.4.1 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Unknown/Multiple Type of failure: None/Unknown | Difficulty: Unknown Test Case: | Blocked By: 4258 Blocking: | Related Tickets: -------------------------------------+------------------------------------ Changes (by tibbe): * status: new => patch -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/5925#comment:17> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

Re: [GHC] #5925: Add inline version of newArray#
by GHC 08 Mar '14

08 Mar '14

#5925: Add inline version of newArray# -------------------------------------+------------------------------------ Reporter: tibbe | Owner: simonmar Type: feature request | Status: new Priority: normal | Milestone: 7.6.2 Component: Compiler | Version: 7.4.1 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Unknown/Multiple Type of failure: None/Unknown | Difficulty: Unknown Test Case: | Blocked By: 4258 Blocking: | Related Tickets: -------------------------------------+------------------------------------ Changes (by tibbe): * owner: => simonmar Comment: Simon, if you could point out any opportunities to break out code from `doNewArrayOp` into reusable allocation primitives that would be great. After this patch I want to support inline allocation in the clone array family of primops, so the more reusable the pieces are the better the code will look. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/5925#comment:16> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

Re: [GHC] #5925: Add inline version of newArray#
by GHC 08 Mar '14

08 Mar '14

#5925: Add inline version of newArray# -------------------------------------+------------------------------------ Reporter: tibbe | Owner: Type: feature request | Status: new Priority: normal | Milestone: 7.6.2 Component: Compiler | Version: 7.4.1 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Unknown/Multiple Type of failure: None/Unknown | Difficulty: Unknown Test Case: | Blocked By: 4258 Blocking: | Related Tickets: -------------------------------------+------------------------------------ Comment (by tibbe): I have something that works. I will post a patch later tonight. Benchmark: allocating 100,000,000 `MutableArray# ()` of size 16 in a loop. Best out of three runs for old implementation: {{{ $ time ./NewArraySlow real 0m2.741s user 0m2.696s sys 0m0.043s }}} Best out of three runs for new implementation: {{{ $ time ./NewArrayNew real 0m1.490s user 0m1.449s sys 0m0.041s }}} That's almost 2x faster. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/5925#comment:15> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 0

[GHC] #8862: forkProcess does not play well with heap or time profiling
by GHC 08 Mar '14

08 Mar '14

#8862: forkProcess does not play well with heap or time profiling --------------------------+------------------------------------------------ Reporter: | Owner: simonmar bennofs | Status: new Type: bug | Milestone: Priority: normal | Version: 7.6.3 Component: | Operating System: Unknown/Multiple Runtime System | Type of failure: Incorrect result at runtime Keywords: | Test Case: Architecture: | Blocking: Unknown/Multiple | Difficulty: | Unknown | Blocked By: | Related Tickets: | --------------------------+------------------------------------------------ This is similar to #4512. When doing heap or time profiling, the forked process and the parent process both write to the same `.hp` or `.prof` file. I think this also applies to program coverage using hpc (didn't test this). I was able to reproduce the bug with the attached source code, but some other people were not. Just run `space-profiling +RTS -h` and try to convert the generated heap profile using `hp2ps`, I get the following error message: {{{ hp2ps: space-profiling.hp, line 186: integer must follow identifier }}} I attached the generated hp file. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8862> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 3

[GHC] #8860: Optimized Cmm isn't
by GHC 08 Mar '14

08 Mar '14

#8860: Optimized Cmm isn't ------------------------------------+------------------------------------- Reporter: tibbe | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.6.3 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- The optimizer seems to miss very basic optimizations. For example, look at this segment (full source below): {{{ _c1Cd::I64 = 0; _c1Ce::I64 = 0; _c1Cf::I64 = 8; call MO_Memset(_c1Cc::I64, _c1Cd::I64, _c1Ce::I64, _c1Cf::I64); }}} I'd expect all these temporaries (that are only used ones) to be inlined, but they are not and thus we fail to see the static arguments to the `MO_Memset` call, which leads to further missed optimizations. This used to work in the old codegen. Furthermore, there are useless basic blocks in the output: {{{ c1Cm: goto c1C4; c1C4: }}} I'd expect them to be eliminated. .dump-cmm file: {{{ ==================== Cmm produced by new codegen ==================== 2014-03-08 08:25:31.728672 UTC [section "data" { a_r1za_closure: const a_r1za_info; }, a_r1za_entry() // [] { info_tbl: [(c1Cm, label: a_r1za_info rep:HeapRep static { Fun {arity: 1 fun_type: ArgSpec 3} })] stack_info: arg_space: 8 updfr_space: Just 8 } {offset c1Cm: goto c1C4; c1C4: if ((old + 0) - <highSp> < SpLim) goto c1Cp; else goto c1Cq; c1Cq: goto c1C3; c1C3: Hp = Hp + 152; if (Hp > HpLim) goto c1Cs; else goto c1Cr; c1Cs: HpAlloc = 152; goto c1Cp; c1Cp: R1 = a_r1za_closure; call (stg_gc_fun)(R1) args: 8, res: 0, upd: 8; c1Cr: I64[Hp - 144] = stg_MUT_ARR_PTRS_DIRTY_info; I64[Hp - 136] = 16; I64[Hp - 128] = 16; _c1C6::I64 = Hp - 144; _c1C7::I64 = _c1C6::I64 + 24; goto c1C8; c1C8: if (_c1C7::I64 < (_c1C6::I64 + 128)) goto c1Ca; else goto c1C9; c1Ca: P64[_c1C7::I64] = ()_closure+1; _c1C7::I64 = _c1C7::I64 + 8; goto c1C8; c1C9: _c1Cb::I64 = _c1C6::I64 + 24; _c1Cc::I64 = _c1Cb::I64 + 128; _c1Cd::I64 = 0; _c1Ce::I64 = 0; _c1Cf::I64 = 8; call MO_Memset(_c1Cc::I64, _c1Cd::I64, _c1Ce::I64, _c1Cf::I64); _s1BY::P64 = _c1C6::I64; _s1BY::P64 = _s1BY::P64; goto c1Ch; c1Ch: _s1C1::P64 = _s1BY::P64; goto c1Ck; c1Ck: R1 = ()_closure+1; call (P64[(old + 8)])(R1) args: 8, res: 0, upd: 8; } }] ==================== Post control-flow optimisations ==================== 2014-03-08 08:25:31.729764 UTC {offset c1Cm: if ((old + 0) - <highSp> < SpLim) goto c1Cp; else goto c1Cq; c1Cq: Hp = Hp + 152; if (Hp > HpLim) goto c1Cs; else goto c1Cr; c1Cs: HpAlloc = 152; goto c1Cp; c1Cp: R1 = a_r1za_closure; call (stg_gc_fun)(R1) args: 8, res: 0, upd: 8; c1Cr: I64[Hp - 144] = stg_MUT_ARR_PTRS_DIRTY_info; I64[Hp - 136] = 16; I64[Hp - 128] = 16; _c1C6::I64 = Hp - 144; _c1C7::I64 = _c1C6::I64 + 24; goto c1C8; c1C8: if (_c1C7::I64 < (_c1C6::I64 + 128)) goto c1Ca; else goto c1C9; c1Ca: P64[_c1C7::I64] = ()_closure+1; _c1C7::I64 = _c1C7::I64 + 8; goto c1C8; c1C9: _c1Cb::I64 = _c1C6::I64 + 24; _c1Cc::I64 = _c1Cb::I64 + 128; _c1Cd::I64 = 0; _c1Ce::I64 = 0; _c1Cf::I64 = 8; call MO_Memset(_c1Cc::I64, _c1Cd::I64, _c1Ce::I64, _c1Cf::I64); _s1BY::P64 = _c1C6::I64; _s1BY::P64 = _s1BY::P64; _s1C1::P64 = _s1BY::P64; R1 = ()_closure+1; call (P64[(old + 8)])(R1) args: 8, res: 0, upd: 8; } ==================== Layout Stack ==================== 2014-03-08 08:25:31.730253 UTC {offset c1Cm: goto c1Cq; c1Cq: Hp = Hp + 152; if (Hp > HpLim) goto c1Cs; else goto c1Cr; c1Cs: HpAlloc = 152; goto c1Cp; c1Cp: R1 = a_r1za_closure; call (stg_gc_fun)(R1) args: 8, res: 0, upd: 8; c1Cr: I64[Hp - 144] = stg_MUT_ARR_PTRS_DIRTY_info; I64[Hp - 136] = 16; I64[Hp - 128] = 16; _c1C6::I64 = Hp - 144; _c1C7::I64 = _c1C6::I64 + 24; goto c1C8; c1C8: if (_c1C7::I64 < (_c1C6::I64 + 128)) goto c1Ca; else goto c1C9; c1Ca: P64[_c1C7::I64] = ()_closure+1; _c1C7::I64 = _c1C7::I64 + 8; goto c1C8; c1C9: _c1Cb::I64 = _c1C6::I64 + 24; _c1Cc::I64 = _c1Cb::I64 + 128; _c1Cd::I64 = 0; _c1Ce::I64 = 0; _c1Cf::I64 = 8; call MO_Memset(_c1Cc::I64, _c1Cd::I64, _c1Ce::I64, _c1Cf::I64); _s1BY::P64 = _c1C6::I64; _s1BY::P64 = _s1BY::P64; _s1C1::P64 = _s1BY::P64; R1 = ()_closure+1; call (P64[Sp])(R1) args: 8, res: 0, upd: 8; } ==================== CAFEnv ==================== 2014-03-08 08:25:31.73074 UTC [(c1C8, {}), (c1C9, {}), (c1Ca, {}), (c1Cm, {}), (c1Cp, {}), (c1Cq, {}), (c1Cr, {}), (c1Cs, {})] ==================== after setInfoTableStackMap ==================== 2014-03-08 08:25:31.730895 UTC a_r1za_entry() // [] { info_tbl: [(c1Cm, label: a_r1za_info rep:HeapRep static { Fun {arity: 1 fun_type: ArgSpec 3} })] stack_info: arg_space: 8 updfr_space: Just 8 } {offset c1Cm: goto c1Cq; c1Cq: Hp = Hp + 152; if (Hp > HpLim) goto c1Cs; else goto c1Cr; c1Cs: HpAlloc = 152; goto c1Cp; c1Cp: R1 = a_r1za_closure; call (stg_gc_fun)(R1) args: 8, res: 0, upd: 8; c1Cr: I64[Hp - 144] = stg_MUT_ARR_PTRS_DIRTY_info; I64[Hp - 136] = 16; I64[Hp - 128] = 16; _c1C6::I64 = Hp - 144; _c1C7::I64 = _c1C6::I64 + 24; goto c1C8; c1C8: if (_c1C7::I64 < (_c1C6::I64 + 128)) goto c1Ca; else goto c1C9; c1Ca: P64[_c1C7::I64] = ()_closure+1; _c1C7::I64 = _c1C7::I64 + 8; goto c1C8; c1C9: _c1Cb::I64 = _c1C6::I64 + 24; _c1Cc::I64 = _c1Cb::I64 + 128; _c1Cd::I64 = 0; _c1Ce::I64 = 0; _c1Cf::I64 = 8; call MO_Memset(_c1Cc::I64, _c1Cd::I64, _c1Ce::I64, _c1Cf::I64); _s1BY::P64 = _c1C6::I64; _s1BY::P64 = _s1BY::P64; _s1C1::P64 = _s1BY::P64; R1 = ()_closure+1; call (P64[Sp])(R1) args: 8, res: 0, upd: 8; } } ==================== Post control-flow optimisations ==================== 2014-03-08 08:25:31.731383 UTC a_r1za_entry() // [] { info_tbl: [(c1Cm, label: a_r1za_info rep:HeapRep static { Fun {arity: 1 fun_type: ArgSpec 3} })] stack_info: arg_space: 8 updfr_space: Just 8 } {offset c1Cm: goto c1Cq; c1Cq: Hp = Hp + 152; if (Hp > HpLim) goto c1Cs; else goto c1Cr; c1Cs: HpAlloc = 152; goto c1Cp; c1Cp: R1 = a_r1za_closure; call (stg_gc_fun)(R1) args: 8, res: 0, upd: 8; c1Cr: I64[Hp - 144] = stg_MUT_ARR_PTRS_DIRTY_info; I64[Hp - 136] = 16; I64[Hp - 128] = 16; _c1C6::I64 = Hp - 144; _c1C7::I64 = _c1C6::I64 + 24; goto c1C8; c1C8: if (_c1C7::I64 < (_c1C6::I64 + 128)) goto c1Ca; else goto c1C9; c1Ca: P64[_c1C7::I64] = ()_closure+1; _c1C7::I64 = _c1C7::I64 + 8; goto c1C8; c1C9: _c1Cb::I64 = _c1C6::I64 + 24; _c1Cc::I64 = _c1Cb::I64 + 128; _c1Cd::I64 = 0; _c1Ce::I64 = 0; _c1Cf::I64 = 8; call MO_Memset(_c1Cc::I64, _c1Cd::I64, _c1Ce::I64, _c1Cf::I64); _s1BY::P64 = _c1C6::I64; _s1BY::P64 = _s1BY::P64; _s1C1::P64 = _s1BY::P64; R1 = ()_closure+1; call (P64[Sp])(R1) args: 8, res: 0, upd: 8; } } ==================== Post CPS Cmm ==================== 2014-03-08 08:25:31.731772 UTC [section "data" { a_r1za_closure: const a_r1za_info; }, a_r1za_entry() // [] { info_tbl: [(c1Cm, label: a_r1za_info rep:HeapRep static { Fun {arity: 1 fun_type: ArgSpec 3} })] stack_info: arg_space: 8 updfr_space: Just 8 } {offset c1Cm: goto c1Cq; c1Cq: Hp = Hp + 152; if (Hp > HpLim) goto c1Cs; else goto c1Cr; c1Cs: HpAlloc = 152; goto c1Cp; c1Cp: R1 = a_r1za_closure; call (stg_gc_fun)(R1) args: 8, res: 0, upd: 8; c1Cr: I64[Hp - 144] = stg_MUT_ARR_PTRS_DIRTY_info; I64[Hp - 136] = 16; I64[Hp - 128] = 16; _c1C6::I64 = Hp - 144; _c1C7::I64 = _c1C6::I64 + 24; goto c1C8; c1C8: if (_c1C7::I64 < (_c1C6::I64 + 128)) goto c1Ca; else goto c1C9; c1Ca: P64[_c1C7::I64] = ()_closure+1; _c1C7::I64 = _c1C7::I64 + 8; goto c1C8; c1C9: _c1Cb::I64 = _c1C6::I64 + 24; _c1Cc::I64 = _c1Cb::I64 + 128; _c1Cd::I64 = 0; _c1Ce::I64 = 0; _c1Cf::I64 = 8; call MO_Memset(_c1Cc::I64, _c1Cd::I64, _c1Ce::I64, _c1Cf::I64); _s1BY::P64 = _c1C6::I64; _s1BY::P64 = _s1BY::P64; _s1C1::P64 = _s1BY::P64; R1 = ()_closure+1; call (P64[Sp])(R1) args: 8, res: 0, upd: 8; } }] ==================== Output Cmm ==================== 2014-03-08 08:25:31.73228 UTC [section "data" { a_r1za_closure: const a_r1za_info; }, a_r1za_entry() // [] { info_tbl: [(c1Cm, label: a_r1za_info rep:HeapRep static { Fun {arity: 1 fun_type: ArgSpec 3} })] stack_info: arg_space: 8 updfr_space: Just 8 } {offset c1Cm: goto c1Cq; c1Cq: Hp = Hp + 152; if (Hp > HpLim) goto c1Cs; else goto c1Cr; c1Cs: HpAlloc = 152; goto c1Cp; c1Cp: R1 = a_r1za_closure; call (stg_gc_fun)(R1) args: 8, res: 0, upd: 8; c1Cr: I64[Hp - 144] = stg_MUT_ARR_PTRS_DIRTY_info; I64[Hp - 136] = 16; I64[Hp - 128] = 16; _c1C6::I64 = Hp - 144; _c1C7::I64 = _c1C6::I64 + 24; goto c1C8; c1C8: if (_c1C7::I64 < (_c1C6::I64 + 128)) goto c1Ca; else goto c1C9; c1Ca: P64[_c1C7::I64] = ()_closure+1; _c1C7::I64 = _c1C7::I64 + 8; goto c1C8; c1C9: _c1Cb::I64 = _c1C6::I64 + 24; _c1Cc::I64 = _c1Cb::I64 + 128; _c1Cd::I64 = 0; _c1Ce::I64 = 0; _c1Cf::I64 = 8; call MO_Memset(_c1Cc::I64, _c1Cd::I64, _c1Ce::I64, _c1Cf::I64); _s1BY::P64 = _c1C6::I64; _s1BY::P64 = _s1BY::P64; _s1C1::P64 = _s1BY::P64; R1 = ()_closure+1; call (P64[Sp])(R1) args: 8, res: 0, upd: 8; } }] }}} .dump-opt-cmm file: {{{ ==================== Optimised Cmm ==================== 2014-03-08 08:25:31.732971 UTC a_r1za_entry() // [] { [(c1Cm, a_r1za_info: const 4294967299; const 0; const 15;)] } {offset c1Cm: goto c1Cq; c1Cq: Hp = Hp + 152; if (Hp > I64[BaseReg + 856]) goto c1Cs; else goto c1Cr; c1Cs: I64[BaseReg + 904] = 152; goto c1Cp; c1Cp: R1 = PicBaseReg + a_r1za_closure; call (I64[BaseReg - 8])(R1) args: 8, res: 0, upd: 8; c1Cr: I64[Hp - 144] = I64[PicBaseReg + stg_MUT_ARR_PTRS_DIRTY_info@GOTPCREL]; I64[Hp - 136] = 16; I64[Hp - 128] = 16; _c1C6::I64 = Hp - 144; _c1C7::I64 = _c1C6::I64 + 24; goto c1C8; c1C8: if (_c1C7::I64 < (_c1C6::I64 + 128)) goto c1Ca; else goto c1C9; c1Ca: I64[_c1C7::I64] = PicBaseReg + (()_closure+1); _c1C7::I64 = _c1C7::I64 + 8; goto c1C8; c1C9: _c1Cb::I64 = _c1C6::I64 + 24; _c1Cc::I64 = _c1Cb::I64 + 128; _c1Cd::I64 = 0; _c1Ce::I64 = 0; _c1Cf::I64 = 8; call MO_Memset(_c1Cc::I64, _c1Cd::I64, _c1Ce::I64, _c1Cf::I64); _s1BY::P64 = _c1C6::I64; // nop _s1C1::P64 = _s1BY::P64; R1 = PicBaseReg + (()_closure+1); call (P64[Sp])(R1) args: 8, res: 0, upd: 8; } } }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8860> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 5