Re: [GHC] #3181: Regression in unboxing

[GHC]

#3181: Regression in unboxing
--+-
  Reporter:  dolio|  Owner: 
  Type:  bug  | Status:  closed 
  Priority:  normal   |  Milestone:  6.12.3 
 Component:  Compiler |Version:  6.10.2 
Resolution:  invalid  |   Keywords:  unboxing boxing
Difficulty:  Unknown  | Os:  Linux  
  Testcase:   |   Architecture:  x86_64 (amd64) 
   Failure:  Runtime performance bug  |  
--+-
Changes (by igloo):

  * status:  infoneeded => closed
  * resolution:  => invalid


Comment:

 No response from submitter, so closing.

-- 
Ticket URL: 
GHC 
The Glasgow Haskell Compiler
___
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

Re: [GHC] #3181: Regression in unboxing

[GHC]

#3181: Regression in unboxing
+---
Reporter:  dolio|Owner: 
Type:  bug  |   Status:  infoneeded 
Priority:  normal   |Milestone:  6.12.3 
   Component:  Compiler |  Version:  6.10.2 
Keywords:  unboxing boxing  |   Difficulty:  Unknown
  Os:  Linux| Testcase: 
Architecture:  x86_64 (amd64)   |  Failure:  Runtime performance bug
+---
Changes (by igloo):

  * status:  new => infoneeded
  * milestone:  6.12.2 => 6.12.3


-- 
Ticket URL: 
GHC 
The Glasgow Haskell Compiler
___
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

Re: [GHC] #3181: Regression in unboxing

[GHC]

#3181: Regression in unboxing
+---
Reporter:  dolio|Owner: 
Type:  bug  |   Status:  new
Priority:  normal   |Milestone:  6.12.2 
   Component:  Compiler |  Version:  6.10.2 
Keywords:  unboxing boxing  |   Difficulty:  Unknown
  Os:  Linux| Testcase: 
Architecture:  x86_64 (amd64)   |  Failure:  Runtime performance bug
+---

Comment(by igloo):

 I can't reproduce this. With 6.8.2:
 {{{
 $ ghc --version
 The Glorious Glasgow Haskell Compilation System, version 6.8.2
 $ ghc -O -c Vector.hs
 $ ghc -O -c StrictPair.hs -ddump-simpl
 }}}
 I get this:
 {{{
   $w$j_sBq =
 \ (w2_sAq :: GHC.Prim.State# GHC.Prim.RealWorld)
   (ww2_sAt :: GHC.Base.Int)
   (ww3_sAw :: GHC.Prim.Int#) ->
   case w_sAC of wild11_auu { GHC.Base.I# x#_auw ->
   case GHC.Prim.<# x#_auw ww3_sAw of wild2_axd {
 GHC.Base.False ->
   case GHC.Prim.==# x#_auw ww3_sAw of wild12_axg {
 GHC.Base.False ->
   case GHC.Prim.writeIntArray#
  @ GHC.Prim.RealWorld arr_avS
 (GHC.Prim.+# ww_sAF 40) x#_auw w2_sAq
   of s'1_awJ { __DEFAULT ->
   (# s'1_awJ, GHC.Base.() #)
   };
 GHC.Base.True ->
   case GHC.Prim.writeIntArray#
  @ GHC.Prim.RealWorld arr_avS
 (GHC.Prim.+# ww_sAF 40) x#_auw w2_sAq
   of s'1_awJ { __DEFAULT ->
   (# s'1_awJ, GHC.Base.() #)
   }
   };
 GHC.Base.True ->
   case GHC.Prim.writeIntArray#
  @ GHC.Prim.RealWorld arr_avS (GHC.Prim.+#
 ww_sAF 40) ww3_sAw w2_sAq
   of s'1_awJ { __DEFAULT ->
   case ww2_sAt of w3_XAV { GHC.Base.I# ww4_XBA ->
   $wa_sBo wild11_auu ww4_XBA ww1_sAJ s'1_awJ
   }
   }
   }
 }}}
 which seems to have the same boxed `Int`. Can you say exactly how you are
 compiling it to get the result you want with 6.8.2 please?

-- 
Ticket URL: 
GHC 
The Glasgow Haskell Compiler
___
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

Re: [GHC] #3181: Regression in unboxing

[GHC]

#3181: Regression in unboxing
--+-
  Reporter:  dolio|  Owner: 
  Type:  bug  | Status:  new
  Priority:  normal   |  Milestone:  6.12.2 
 Component:  Compiler |Version:  6.10.2 
Resolution:   |   Keywords:  unboxing boxing
Difficulty:  Unknown  | Os:  Linux  
  Testcase:   |   Architecture:  x86_64 (amd64) 
   Failure:  Runtime performance bug  |  
--+-
Changes (by igloo):

  * milestone:  6.12.1 => 6.12.2

-- 
Ticket URL: 
GHC 
The Glasgow Haskell Compiler___
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

Re: [GHC] #3181: Regression in unboxing

[GHC]

#3181: Regression in unboxing
-+--
Reporter:  dolio |Owner: 
Type:  run-time performance bug  |   Status:  new
Priority:  normal|Milestone:  6.12.1 
   Component:  Compiler  |  Version:  6.10.2 
Severity:  normal|   Resolution: 
Keywords:  unboxing boxing   |   Difficulty:  Unknown
Testcase:|   Os:  Linux  
Architecture:  x86_64 (amd64)|  
-+--
Comment (by igloo):

 Thanks for all your work in boiling down the testcase.

-- 
Ticket URL: 
GHC 
The Glasgow Haskell Compiler___
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

Re: [GHC] #3181: Regression in unboxing

[GHC]

#3181: Regression in unboxing
-+--
Reporter:  dolio |Owner: 
Type:  run-time performance bug  |   Status:  new
Priority:  normal|Milestone:  6.12.1 
   Component:  Compiler  |  Version:  6.10.2 
Severity:  normal|   Resolution: 
Keywords:  unboxing boxing   |   Difficulty:  Unknown
Testcase:|   Os:  Linux  
Architecture:  x86_64 (amd64)|  
-+--
Changes (by igloo):

  * milestone:  => 6.12.1

-- 
Ticket URL: 
GHC 
The Glasgow Haskell Compiler___
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

Re: [GHC] #3181: Regression in unboxing

[GHC]

#3181: Regression in unboxing
-+--
Reporter:  dolio |Owner: 
Type:  run-time performance bug  |   Status:  new
Priority:  normal|Milestone: 
   Component:  Compiler  |  Version:  6.10.2 
Severity:  normal|   Resolution: 
Keywords:  unboxing boxing   |   Difficulty:  Unknown
Testcase:|   Os:  Linux  
Architecture:  x86_64 (amd64)|  
-+--
Changes (by pumpkin):

 * cc: pumpkin...@gmail.com (added)

-- 
Ticket URL: 
GHC 
The Glasgow Haskell Compiler___
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

Re: [GHC] #3181: Regression in unboxing

[GHC]

#3181: Regression in unboxing
-+--
Reporter:  dolio |Owner: 
Type:  run-time performance bug  |   Status:  new
Priority:  normal|Milestone: 
   Component:  Compiler  |  Version:  6.10.2 
Severity:  normal|   Resolution: 
Keywords:  unboxing boxing   |   Difficulty:  Unknown
Testcase:|   Os:  Linux  
Architecture:  x86_64 (amd64)|  
-+--
Comment (by dolio):

 I took the plunge and implemented a very reduced version of
 Data.Array.Vector (should be attached now).

 It appears to generate similar code to me, and it does result in the
 unboxing bug, so I think this should be self-contained now. Obviously
 Vector.hs needs to go in Data/Array/ below StrictPair.hs

 Hope that helps.

-- 
Ticket URL: 
GHC 
The Glasgow Haskell Compiler___
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

Re: [GHC] #3181: Regression in unboxing

[GHC]

#3181: Regression in unboxing
-+--
Reporter:  dolio |Owner: 
Type:  run-time performance bug  |   Status:  new
Priority:  normal|Milestone: 
   Component:  Compiler  |  Version:  6.10.2 
Severity:  normal|   Resolution: 
Keywords:  unboxing boxing   |   Difficulty:  Unknown
Testcase:|   Os:  Linux  
Architecture:  x86_64 (amd64)|  
-+--
Comment (by simonpj):

 OK that's good enough I guess.  I assume that 0.1.0.3 will stay there even
 if later versions get uploaded.  Thx

-- 
Ticket URL: 
GHC 
The Glasgow Haskell Compiler___
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

Re: [GHC] #3181: Regression in unboxing

[GHC]

#3181: Regression in unboxing
-+--
Reporter:  dolio |Owner: 
Type:  run-time performance bug  |   Status:  new
Priority:  normal|Milestone: 
   Component:  Compiler  |  Version:  6.10.2 
Severity:  normal|   Resolution: 
Keywords:  unboxing boxing   |   Difficulty:  Unknown
Testcase:|   Os:  Linux  
Architecture:  x86_64 (amd64)|  
-+--
Comment (by dolio):

 I was using uvector-0.2, which is only available via darcs currently.
 However, I just tested with 0.1.0.3 (currently the latest) from hackage
 and get the same results.

-- 
Ticket URL: 
GHC 
The Glasgow Haskell Compiler___
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

Re: [GHC] #3181: Regression in unboxing

[GHC]

#3181: Regression in unboxing
-+--
Reporter:  dolio |Owner: 
Type:  run-time performance bug  |   Status:  new
Priority:  normal|Milestone: 
   Component:  Compiler  |  Version:  6.10.2 
Severity:  normal|   Resolution: 
Keywords:  unboxing boxing   |   Difficulty:  Unknown
Testcase:|   Os:  Linux  
Architecture:  x86_64 (amd64)|  
-+--
Comment (by simonpj):

 Not unacceptable, just less convenient.  If you've failed to cut it down
 with a modest investment, then don't worry.

 However, can you pls say exactly which version of uvector you were using?
 That at least helps to ensure the bug is remains reproducible.

 Simon

-- 
Ticket URL: 
GHC 
The Glasgow Haskell Compiler___
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

Re: [GHC] #3181: Regression in unboxing

[GHC]

#3181: Regression in unboxing
-+--
Reporter:  dolio |Owner: 
Type:  run-time performance bug  |   Status:  new
Priority:  normal|Milestone: 
   Component:  Compiler  |  Version:  6.10.2 
Severity:  normal|   Resolution: 
Keywords:  unboxing boxing   |   Difficulty:  Unknown
Testcase:|   Os:  Linux  
Architecture:  x86_64 (amd64)|  
-+--
Comment (by dolio):

 Ah, sorry. Data.Array.Vector is from the uvector package, which is
 available on hackage. If that's unacceptable, I can try to construct a new
 test case, or make a dummy module as you said, but my initial attempts
 writing do-nothing loops similar to the one above didn't trigger the bug
 (so obviously results are quite variable).

-- 
Ticket URL: 
GHC 
The Glasgow Haskell Compiler___
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

Re: [GHC] #3181: Regression in unboxing

[GHC]

#3181: Regression in unboxing
-+--
Reporter:  dolio |Owner: 
Type:  run-time performance bug  |   Status:  new
Priority:  normal|Milestone: 
   Component:  Compiler  |  Version:  6.10.2 
Severity:  normal|   Resolution: 
Keywords:  unboxing boxing   |   Difficulty:  Unknown
Testcase:|   Os:  Linux  
Architecture:  x86_64 (amd64)|  
-+--
Changes (by simonpj):

  * difficulty:  => Unknown

Comment:

 Thanks for a nice well-characterised bug report.

 I can't reproduce it because you didn't include `Data.Vector`. Could you
 do so?  (It's v helpful for tickets to be self contained where possible.)
 If it's big, you may be able to cut it down to just what the test case
 needs.  I don't necessarily even need to run it: just to compile it.  So
 you may be able to go
 {{{
 module Vector( foo ) where
   foo :: Int -> Int
   foo = error "stub"
 }}}
 or stuff like that.

 Anyway, I have seen cases like this before.  Here's the relevant comment
 from `Simplify.lhs`:
 {{{
 Note [Case binders and join points]
 ~~~
 Consider this
case (case .. ) of c {
  I# c# -> c

 If we make a join point with c but not c# we get
   $j = \c -> c

 But if later inlining scrutines the c, thus

   $j = \c -> ... case c of { I# y -> ... } ...

 we won't see that 'c' has already been scrutinised.
 An alternative plan is this:

$j = \c# -> let c = I# c# in ...c

 but that is bad if 'c' is *not* later scrutinised.
 }}}
 Whether or not it happens depends delicately on exactly when inlining
 takes place. Hence (I think) the regression.

 I have a fix for my test HEAD, but I'd like to use your example as a test
 case.

 Simon

-- 
Ticket URL: 
GHC 
The Glasgow Haskell Compiler___
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs