#9583: Simplifier ticks exhausted while compiling Cabal HEAD -------------------------------------+------------------------------------- Reporter: ezyang | Owner: Type: bug | Status: new Priority: highest | Milestone: Component: Compiler | Version: 7.9 Resolution: | Keywords: Operating System: | Architecture: Unknown/Multiple Unknown/Multiple | Difficulty: Unknown Type of failure: Compile- | Blocked By: time crash | Related Tickets: Test Case: | Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- Comment (by ezyang): Sorry, use the GitHub repo: {{{ git clone http://github.com/haskell/cabal.git }}} It also fails at tick factor 8192, and ran out of memory for me on the next power of two. Probably an infinite loop. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9583#comment:2 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9583: Simplifier ticks exhausted while compiling Cabal HEAD -------------------------------------+------------------------------------- Reporter: ezyang | Owner: Type: bug | Status: new Priority: highest | Milestone: Component: Compiler | Version: 7.9 Resolution: | Keywords: Operating System: | Architecture: Unknown/Multiple Unknown/Multiple | Difficulty: Unknown Type of failure: Compile- | Blocked By: time crash | Related Tickets: Test Case: | Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- Comment (by slyfox): I wonder if #9565 is of the same cause (smaller source file, but nastier). -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9583#comment:4 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9583: Simplifier ticks exhausted while compiling Cabal HEAD -------------------------------------+------------------------------------- Reporter: ezyang | Owner: Type: bug | Status: new Priority: highest | Milestone: Component: Compiler | Version: 7.9 Resolution: | Keywords: Operating System: | Architecture: Unknown/Multiple Unknown/Multiple | Difficulty: Unknown Type of failure: Compile- | Blocked By: time crash | Related Tickets: Test Case: | Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- Comment (by JohnWiegley): Simon and I had a look at this, and we found a few more data points: 1. Manually implementing {{{get}}} and {{{put}}} for the instance, even if just to copy the default definitions, causes the test case above to pass. 2. The following data type also reproduces the problem: {{{ data T = A T T T T T T T T T T T T | B T deriving (Data, Typeable, Generic) }}} 3. If you implement {{{get}}} for the instance, but not {{{put}}}, the error still occurs; but the converse works. 4. If you set {{{put = gdmput}}}, and then define it in its own module, marking it INLINE works, but INLINEABLE fails: {{{ gdmput :: (Generic t, GBinary (Rep t)) => t -> Put gdmput = gput . from {-# INLINABLE gdmput #-} }}} 5. Displaying inlining information with {{{-ddump-inlinings -ddump-rule- firings}}} shows that after the first few unfoldings, the following repeats continually up until the failure. Increasing the tick cause simply causes more of this occurrence to display: {{{ Rule fired: Class op gput Rule fired: Class op gput Rule fired: Class op gput Rule fired: Class op gput Inlining done: Data.Binary.Generic.unTagged Inlining done: Data.Binary.Generic.unTagged1 Inlining done: GHC.Base.id Inlining done: GHC.Word.$fOrdWord64_$c<= Inlining done: GHC.Word.$fNumWord64_$c- Inlining done: Data.Binary.Generic.$fSumSizeM1_$csumSize Inlining done: Data.Binary.Generic.$fSumSizeM1_$csumSize Rule fired: plusWord# Inlining done: GHC.Word.$fBitsWord64_$cfromInteger Rule fired: integerToWord Rule fired: minusWord# Rule fired: leWord# Rule fired: tagToEnum# Inlining done: GHC.Word.$fBitsWord8_$cfromInteger Rule fired: integerToWord Rule fired: narrow8Word# Inlining done: Data.Binary.Generic.$fSumSizeM1_$csumSize Inlining done: Data.Binary.Generic.$fSumSizeM1_$csumSize Rule fired: plusWord# Inlining done: GHC.Base.id Rule fired: narrow8Word# Inlining done: GHC.Word.$fBitsWord8_$cshiftR Rule fired: >=# Rule fired: tagToEnum# Rule fired: uncheckedShiftRL# Rule fired: Class op putSum Inlining done: Data.Binary.Put.$fApplicativePutM_$c*> Inlining done: Data.Binary.Put.$fApplicativePutM2 Rule fired: Class op put Inlining done: Data.Binary.Put.putWord8 Inlining done: GHC.Base.$ Inlining done: GHC.Base.returnIO Inlining done: GHC.Base.returnIO1 Inlining done: GHC.Base.bindIO Inlining done: GHC.Base.bindIO1 Inlining done: GHC.Base.flip Inlining done: Foreign.Storable.$fStorableWord8_$cpoke Inlining done: Foreign.Storable.$fStorableWord19 Rule fired: Class op gput Inlining done: GHC.Word.$fNumWord8_$c+ Rule fired: plusWord# Rule fired: narrow8Word# Inlining done: GHC.Word.$fNumWord8_$c- Rule fired: minusWord# Rule fired: narrow8Word# Rule fired: Class op putSum Inlining done: Data.Binary.Put.$fApplicativePutM_$c*> Inlining done: Data.Binary.Put.$fApplicativePutM2 Rule fired: Class op put Inlining done: Data.Binary.Put.putWord8 Inlining done: GHC.Base.$ Inlining done: GHC.Base.returnIO Inlining done: GHC.Base.returnIO1 Inlining done: GHC.Base.bindIO Inlining done: GHC.Base.bindIO1 Inlining done: GHC.Base.flip Inlining done: Foreign.Storable.$fStorableWord8_$cpoke Inlining done: Foreign.Storable.$fStorableWord19 }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9583#comment:6 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9583: Simplifier ticks exhausted while compiling Cabal HEAD -------------------------------------+------------------------------------- Reporter: ezyang | Owner: Type: bug | Status: new Priority: highest | Milestone: Component: Compiler | Version: 7.9 Resolution: | Keywords: Operating System: | Architecture: Unknown/Multiple Unknown/Multiple | Difficulty: Unknown Type of failure: Compile- | Blocked By: time crash | Related Tickets: Test Case: | Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- Comment (by ezyang): Bisection finished, the bad commit is this one: {{{ b9e49d3e9580e13d89efd1f779cb76f610e0d6e0 is the first bad commit [4/1912] commit b9e49d3e9580e13d89efd1f779cb76f610e0d6e0 Author: Simon Peyton Jones Date: Tue May 13 13:10:26 2014 +0100 Add -fspecialise-aggressively This flag specialises any imported overloaded function that has an unfolding, whether or not it was marked INLINEABLE. We get a lot of orphan SPEC rules as a result, but that doesn't matter provided we don't treat orphan auto-generated rules as causing the module itself to be an orphan module. See Note [Orphans and auto-generated rules] in MkIface. }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9583#comment:8 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9583: Simplifier ticks exhausted while compiling Cabal HEAD -------------------------------------+------------------------------------- Reporter: ezyang | Owner: Type: bug | Status: new Priority: highest | Milestone: Component: Compiler | Version: 7.9 Resolution: | Keywords: Operating System: | Architecture: Unknown/Multiple Unknown/Multiple | Difficulty: Unknown Type of failure: Compile- | Blocked By: time crash | Related Tickets: Test Case: | Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- Comment (by hvr): Replying to [comment:9 simonpj]:

Patch coming

any ETA on that patch? It's blocking a few things right now... :-) -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9583#comment:10 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9583: Simplifier ticks exhausted while compiling Cabal HEAD -------------------------------------+------------------------------------- Reporter: ezyang | Owner: Type: bug | Status: new Priority: highest | Milestone: Component: Compiler | Version: 7.9 Resolution: | Keywords: Operating System: | Architecture: Unknown/Multiple Unknown/Multiple | Difficulty: Unknown Type of failure: Compile- | Blocked By: time crash | Related Tickets: Test Case: | Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- Comment (by simonpj): I've just pushed it. I'm getting a failure on `perf/compiler/T783`, but as usual I'm uncertain about whether the patch is the culprit or not. I suspect not, so since it's a blocker I'm pushing it anyway. Simon -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9583#comment:12 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9583: Simplifier ticks exhausted while compiling Cabal HEAD -------------------------------------+------------------------------------- Reporter: ezyang | Owner: Type: bug | Status: new Priority: highest | Milestone: Component: Compiler | Version: 7.9 Resolution: | Keywords: Operating System: | Architecture: Unknown/Multiple Unknown/Multiple | Difficulty: Unknown Type of failure: Compile- | Blocked By: time crash | Related Tickets: Test Case: | Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- Changes (by simonpj): * cc: dreixel (added) Comment: Perhaps someone can confirm that Cabal is ok, and then close. Pedro: the "deriving Binary" stuff still generates a gargantuan quantity of code! Simon -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9583#comment:14 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9583: Simplifier ticks exhausted while compiling Cabal HEAD -------------------------------------+------------------------------------- Reporter: ezyang | Owner: Type: bug | Status: new Priority: highest | Milestone: Component: Compiler | Version: 7.9 Resolution: | Keywords: Operating System: | Architecture: Unknown/Multiple Unknown/Multiple | Difficulty: Unknown Type of failure: Compile- | Blocked By: time crash | Related Tickets: Test Case: | Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- Comment (by hvr): Replying to [comment:14 simonpj]:

Perhaps someone can confirm that Cabal is ok, and then close.

Pedro: the "deriving Binary" stuff still generates a gargantuan quantity of code!

Compiling Cabal within the GHC tree is not really ok (see comments in Phab:D183). It does compile, but it takes *several* minutes as well as massive amount of memory (over 50% of the 4GiB mem available in the buildbot, causing it to fail w/ an out-of-mem exception since we build mit CPUS=4) to when trying to produce the `libraries/Cabal/Cabal/dist- install/build/Language/Haskell/Extension.o` module with `ghc-stage1` -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9583#comment:16 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9583: Simplifier ticks exhausted while compiling Cabal HEAD -------------------------------------+------------------------------------- Reporter: ezyang | Owner: Type: bug | Status: new Priority: highest | Milestone: Component: Compiler | Version: 7.9 Resolution: | Keywords: Operating System: | Architecture: Unknown/Multiple Unknown/Multiple | Difficulty: Unknown Type of failure: Compile- | Blocked By: time crash | Related Tickets: #9630 Test Case: | Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- Changes (by hvr): * related: => #9630 Comment: I've created a new ticket for the performance regression over at #9630 to separate concerns a bit -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9583#comment:18 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler