This reverts commit e1e68550 which was
accidentally pushed.

Simon Marlow authored 13 years ago and pcapriotti committed 12 years ago

-package P means "the latest version of P" if multiple versions are
installed.  It was working as advertised, but we were
eagerly *linking* all versions of P, which might cause an error if the
package has some C code, because we can't link multiple instances of
the same symbol.

MERGED from commit 62164cf5

td123 authored 13 years ago and pcapriotti committed 13 years ago

Fixes http://hackage.haskell.org/trac/ghc/ticket/5992

Simon Marlow authored 13 years ago and pcapriotti committed 13 years ago

There was also a bug caused by INLINEs getting SCCs, but unfortunately
I have lost the test case.  The Note in the code describes the problem
though.

MERGED from commit 6181e007

Patch by Samuel Thibault.

See http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=659530.

MERGED from commit d2b538c8

Simon Marlow authored 13 years ago and pcapriotti committed 13 years ago

We were keeping around the Task struct (216 bytes) for every worker we
ever created, even though we only keep a maximum of 6 workers per
Capability.  These Task structs accumulate and cause a space leak in
programs that do lots of safe FFI calls; this patch frees the Task
struct as soon as a worker exits.

One reason we were keeping the Task structs around is because we print
out per-Task timing stats in +RTS -s, but that isn't terribly useful.
What is sometimes useful is knowing how *many* Tasks there were.  So
now I'm printing a single-line summary, this is for the program in

  TASKS: 2001 (1 bound, 31 peak workers (2000 total), using -N1)

So although we created 2k tasks overall, there were only 31 workers
active at any one time (which is exactly what we expect: the program
makes 30 safe FFI calls concurrently).

This also gives an indication of how many capabilities were being
used, which is handy if you use +RTS -N without an explicit number.

MERGED from commit 085c7fe5

Simon Marlow authored 13 years ago and pcapriotti committed 13 years ago

MERGED from commit 20ba7f1a

Simon Marlow authored 13 years ago and pcapriotti committed 13 years ago

A thunk with no free variables was not getting blackholed when
-feager-blackholing was on, but we were nevertheless pushing the
stg_bh_upd_frame version of the update frame that expects to see a
black hole.

I fixed this twice for good measure:

 - we now call blackHoleOnEntry when pushing the update frame to check
   whether the closure was actually blackholed, and so that we use the
   same predicate in both places

 - we now black hole thunks even if they have no free variables.
   These only occur when optimisation is off, but presumably if you say
   -feager-blackholing then that's what you want to happen.

MERGED from commit 21a53a1c

Simon Marlow authored 13 years ago and pcapriotti committed 13 years ago

MERGED from commit d82da293

Simon Marlow authored 13 years ago and pcapriotti committed 13 years ago

MERGED from commit c95342ce

Simon Marlow authored 13 years ago and pcapriotti committed 13 years ago

Hopefully should fix random "bad heap profile" failures in the nightly
builds.

MERGED from commit 634fb8d3

Merge of:
    commit 4fa3f16d
    Author: Simon Peyton Jones <simonpj@microsoft.com>
    Date:   Mon May 28 17:33:42 2012 +0100

Be less aggressive about the result discount

This patch fixes Trac #6099 by reducing the result discount in CoreUnfold.conSize.
See Note [Constructor size and result discount] in CoreUnfold.

The existing version is definitely too aggressive. Simon M found it an
"unambiguous win" but it is definitely what led to the bloat. In a function
with a lot of case branches, all returning a constructor, the discount could
grow arbitrarily large.

I also had to increase the -funfolding-creation-threshold from 450 to 750,
otherwise some functions that should inline simply never get an unfolding.
(The massive result discount was allow the unfolding to appear before.)

The nofib results are these, picking a handful of outliers to show.

        Program           Size    Allocs   Runtime   Elapsed  TotalMem
--------------------------------------------------------------------------------
         fulsom          -0.5%     -1.6%     -2.8%     -2.6%    +31.1%
       maillist          -0.2%     -0.0%      0.09      0.09     -3.7%
         mandel          -0.4%     +6.6%      0.12      0.12     +0.0%
       nucleic2          -0.2%    +18.5%      0.11      0.11     +0.0%
        parstof          -0.4%     +4.0%      0.00      0.00     +0.0%
--------------------------------------------------------------------------------
            Min          -0.9%     -1.6%    -19.7%    -19.7%     -3.7%
            Max          +0.3%    +18.5%     +2.7%     +2.7%    +31.1%
 Geometric Mean          -0.3%     +0.4%     -3.0%     -3.0%     +0.2%

Turns out that nucleic2 has a function
  Main.$wabsolute_pos =
    \ (ww_s4oj :: Types.Tfo) (ww1_s4oo :: Types.FloatT)
      (ww2_s4op :: Types.FloatT) (ww3_s4oq :: Types.FloatT) ->
      case ww_s4oj
      of _
      { Types.Tfo a_a1sS b_a1sT c_a1sU d_a1sV e_a1sW f_a1sX g_a1sY h_a1sZ i_a1t0 tx_a1t1 ty_a1t2 tz_a1t3 ->
      (# case ww1_s4oo of _ { GHC.Types.F# x_a2sO ->
         case a_a1sS of _ { GHC.Types.F# y_a2sS ->
         case ww2_s4op of _ { GHC.Types.F# x1_X2y9 ->
         case d_a1sV of _ { GHC.Types.F# y1_X2yh ->
         case ww3_s4oq of _ { GHC.Types.F# x2_X2yj ->
         case g_a1sY of _ { GHC.Types.F# y2_X2yr ->
         case tx_a1t1 of _ { GHC.Types.F# y3_X2yn ->
         GHC.Types.F#
           (GHC.Prim.plusFloat#
              (GHC.Prim.plusFloat#
                 (GHC.Prim.plusFloat#
                    (GHC.Prim.timesFloat# x_a2sO y_a2sS)
                    (GHC.Prim.timesFloat# x1_X2y9 y1_X2yh))
                 (GHC.Prim.timesFloat# x2_X2yj y2_X2yr))
              y3_X2yn)
         } } }}}}},

        <similar>,
        <similar> )

This is pretty big, but inlining it does get rid of that F# allocation.
But we'll also get rid of it with deep CPR: Trac #2289. For now we just
accept the change.

Simon Marlow authored 13 years ago and pcapriotti committed 13 years ago

MERGED from commit b6e28266

Simon Marlow authored 13 years ago and pcapriotti committed 13 years ago

MERGED from commit 4ca28182

To prevent conflicts with GCC builtins, generate identical code for
calls to mem primos and FFI calls.

Based on a patch by Joachim Breitner.
MERGED from commit 40c1106c

The reasoning is that GHC will only skip recompilation if it has found
a valid up-to-date object file and .hi file, so there is no need to
also check that these flags have not changed, and indeed there are
valid reasons to want to change them.

Fixes gas errors when -g is enabled:

Error: can't resolve `.debug-ghc-link-info' {.debug-ghc-link-info section} - `.Ltext0' {.text section}

MERGED from commit 18c2a2f7

MERGED from commit 44098b3d

Simon Peyton Jones authored 13 years ago and pcapriotti committed 13 years ago

A long-standing and egregious bug in the worker/wrapper code meant
that some functions with the CPR property weren't getting a CPR
w/w. And that had the effect of making a tail-recursive function not
tail recursive.  As well as increasing allocation.

Fixes Trac #5920, and #5997.

Nofib results (highlights):

        Program           Size    Allocs   Runtime   Elapsed  TotalMem
--------------------------------------------------------------------------------
         boyer2          -0.1%    -15.3%      0.01      0.01     +0.0%
        mandel2          -0.0%     -8.1%      0.01      0.01     +0.0%
           para          -0.1%    -11.8%     -7.9%     -7.8%     +0.0%
--------------------------------------------------------------------------------
            Min          -0.1%    -15.3%     -7.9%     -7.8%    -33.3%
            Max          +0.0%     +0.2%     +6.3%     +6.3%     +3.7%
 Geometric Mean          -0.0%     -0.4%     +0.1%     +0.1%     -0.5%

Looks like a clear win.  And I have not even recompiled the libraries, so
it'll probably be a bit better in the ed.

MERGED from commit b8ff4448

Simon Marlow authored 13 years ago and pcapriotti committed 13 years ago

MERGED from commit 5eabdc16

Ian Lynagh authored 13 years ago and pcapriotti committed 13 years ago

I don't think it was causing any problems, but
    TimeToUS(x+y)
would have evaluated to
    x + (y / 1000)

MERGED from commit 54121fff

On the 7.4 branch, this was causing a wrong conversion in
getDelayTarget.

Signed-off-by: Paolo Capriotti <p.capriotti@gmail.com>

Simon Marlow authored 13 years ago and pcapriotti committed 13 years ago

The threaded RTS had a fix for this a long time ago (#1623) but this
patch applies a similar fix to the non-threaded RTS.

MERGED from commit dd24d6bc

Simon Marlow authored 13 years ago and pcapriotti committed 13 years ago

Firstly, we were rounding up too much, such that the smallest delay
was 20ms.  Secondly, there is no need to use millisecond resolution on
a 64-bit machine where we have room in the TSO to use the normal
nanosecond resolution that we use elsewhere in the RTS.

MERGED from commit 9d26519c

David Terei authored 13 years ago and pcapriotti committed 13 years ago

This fixes haddock so it correctly reports
the safe haskell mode of a module.

MERGED from commit e7e5e277

Simon Peyton Jones authored 13 years ago and pcapriotti committed 13 years ago

This allows us to give a sensible error message when a Template Haskell
splice generates an occurrence without a binding site.

Fixes Trac #5959 and #5971

MERGED from commit cb5a3f2d

This reverts commit 6a217ed6.