1. 12 Jun, 2019 1 commit
  2. 30 May, 2019 1 commit
  3. 26 May, 2019 1 commit
    • isovector's avatar
      Let the specialiser work on dicts under lambdas · 2d0cf625
      isovector authored and  Marge Bot's avatar Marge Bot committed
      Following the discussion under #16473, this change allows the
      specializer to work on any dicts in a lambda, not just those that occur
      at the beginning.
      
      For example, if you use data types which contain dictionaries and
      higher-rank functions then once these are erased by the optimiser you
      end up with functions such as:
      
      ```
        go_s4K9
        Int#
        -> forall (m :: * -> *).
           Monad m =>
           (forall x. Union '[State (Sum Int)] x -> m x) -> m ()
      ```
      
      The dictionary argument is after the Int# value argument, this patch
      allows `go` to be specialised.
      2d0cf625
  4. 01 May, 2019 1 commit
    • Sebastian Graf's avatar
      Compute demand signatures assuming idArity · 014ed644
      Sebastian Graf authored and  Marge Bot's avatar Marge Bot committed
      This does four things:
      
      1. Look at `idArity` instead of manifest lambdas to decide whether to use LetUp
      2. Compute the strictness signature in LetDown assuming at least `idArity`
         incoming arguments
      3. Remove the special case for trivial RHSs, which is subsumed by 2
      4. Don't perform the W/W split when doing so would eta expand a binding.
         Otherwise we would eta expand PAPs, causing unnecessary churn in the
         Simplifier.
      
      NoFib Results
      
      --------------------------------------------------------------------------------
              Program         Allocs    Instrs
      --------------------------------------------------------------------------------
       fannkuch-redux          +0.3%      0.0%
                   gg          -0.0%     -0.1%
             maillist          +0.2%     +0.2%
              minimax           0.0%     +0.8%
               pretty           0.0%     -0.1%
              reptile          -0.0%     -1.2%
      --------------------------------------------------------------------------------
                  Min          -0.0%     -1.2%
                  Max          +0.3%     +0.8%
       Geometric Mean          +0.0%     -0.0%
      014ed644
  5. 08 Apr, 2019 1 commit
  6. 20 Mar, 2019 1 commit
  7. 08 Mar, 2019 1 commit
  8. 24 Feb, 2019 1 commit
  9. 14 Feb, 2019 1 commit
  10. 30 Jan, 2019 3 commits
  11. 07 Nov, 2018 1 commit
    • davide's avatar
      testsuite: Save performance metrics in git notes. · 932cd41d
      davide authored and Ben Gamari's avatar Ben Gamari committed
      This patch makes the following improvement:
        - Automatically records test metrics (per test environment) so that
          the programmer need not supply nor update expected values in *.T
          files.
          - On expected metric changes, the programmer need only indicate the
            direction of change in the git commit message.
        - Provides a simple python tool "perf_notes.py" to compare metrics
          over time.
      
      Issues:
        - Using just the previous commit allows performance to drift with each
          commit.
          - Currently we allow drift as we have a preference for minimizing
            false positives.
          - Some possible alternatives include:
            - Use metrics from a fixed commit per test: the last commit that
              allowed a change in performance (else the oldest metric)
            - Or use some sort of aggregate since the last commit that allowed
              a change in performance (else all available metrics)
            - These alternatives may result in a performance issue (with the
              test driver) having to heavily search git commits/notes.
        - Run locally, performance tests will trivially pass unless the tests
          were run locally on the previous commit. This is often not the case
          e.g.  after pulling recent changes.
      
      Previously, *.T files contain statements such as:
      ```
      stats_num_field('peak_megabytes_allocated', (2, 1))
      compiler_stats_num_field('bytes allocated',
                               [(wordsize(64), 165890392, 10)])
      ```
      This required the programmer to give the expected values and a tolerance
      deviation (percentage). With this patch, the above statements are
      replaced with:
      ```
      collect_stats('peak_megabytes_allocated', 5)
      collect_compiler_stats('bytes allocated', 10)
      ```
      So that programmer must only enter which metrics to test and a tolerance
      deviation. No expected value is required. CircleCI will then run the
      tests per test environment and record the metrics to a git note for that
      commit and push them to the git.haskell.org ghc repo. Metrics will be
      compared to the previous commit. If they are different by the tolerance
      deviation from the *.T file, then the corresponding test will fail. By
      adding to the git commit message e.g.
      ```
       # Metric (In|De)crease <metric(s)> <options>: <tests>
      Metric Increase ['bytes allocated', 'peak_megabytes_allocated'] \
               (test_env='linux_x86', way='default'):
          Test012, Test345
      Metric Decrease 'bytes allocated':
          Test678
      Metric Increase:
          Test711
      ```
      This will allow the noted changes (letting the test pass). Note that by
      omitting metrics or options, the change will apply to all possible
      metrics/options (i.e. in the above, an increase for all metrics in all
      test environments is allowed for Test711)
      
      phabricator will use the message in the description
      
      Reviewers: bgamari, hvr
      
      Reviewed By: bgamari
      
      Subscribers: rwbarton, carter
      
      GHC Trac Issues: #12758
      
      Differential Revision: https://phabricator.haskell.org/D5059
      932cd41d
  12. 05 Oct, 2018 1 commit
  13. 22 Sep, 2018 1 commit
  14. 21 Sep, 2018 2 commits
  15. 14 Sep, 2018 1 commit
  16. 06 Aug, 2018 1 commit
  17. 04 Aug, 2018 1 commit
  18. 27 Jul, 2018 1 commit
    • Simon Peyton Jones's avatar
      Refactor (~) to reduce the suerpclass stack · f265008f
      Simon Peyton Jones authored
      The constraint (~) used to be (effectively):
        class a ~~ b => (a :: k) ~ (b :: k)
      
      but, with this patch, it is now defined uniformly with
      (~~) and Coercible like this:
        class a ~# b => (a :: k) ~ (b :: k)
      
      Result:
        * One less superclass selection when goinng from (~) to (~#)
          Better for compile time and better for debugging with -ddump-simpl
      
        * The code for (~), (~~), and Coercible looks uniform, and appears
          together, e.g. in TysWiredIn and ClsInst.matchGlobalInst.
          Previously the code for (~) was different, and unique.
      
      Not only is this simpler, but it also makes the compiler a bit faster;
        T12227: 9% less allocation
        T12545: 7% less allocation
      
      This patch fixes Trac #15421
      f265008f
  19. 16 Jul, 2018 1 commit
    • Simon Marlow's avatar
      Fix space leaks · 71f6b18b
      Simon Marlow authored
      Summary:
      All these were detected by -fghci-leak-check when GHC was
      compiled *without* optimisation (e.g. using the "quick" build flavour).
      
      Unfortunately I don't know of a good way to keep this working.  I'd like
      to just disable the -fghci-leak-check flag when the compiler is built
      without optimisation, but it doesn't look like we have an easy way to do
      that. And even if we could, it would be fragile anyway,
      
      Test Plan: `cd testsuite/tests/ghci; make`
      
      Reviewers: bgamari, hvr, erikd, tdammers
      
      Subscribers: tdammers, rwbarton, thomie, carter
      
      GHC Trac Issues: #15246
      
      Differential Revision: https://phabricator.haskell.org/D4872
      71f6b18b
  20. 10 Jul, 2018 1 commit
    • Ningning Xie's avatar
      Refactor coercion rule · 55a3f855
      Ningning Xie authored and Richard Eisenberg's avatar Richard Eisenberg committed
      Summary:
      The patch is an attempt on #15192.
      
      It defines a new coercion rule
      
      ```
       | GRefl Role Type MCoercion
      ```
      
      which correspondes to the typing rule
      
      ```
           t1 : k1
        ------------------------------------
        GRefl r t1 MRefl: t1 ~r t1
      
           t1 : k1       co :: k1 ~ k2
        ------------------------------------
        GRefl r t1 (MCo co) : t1 ~r t1 |> co
      ```
      
      MCoercion wraps a coercion, which might be reflexive (MRefl)
      or not (MCo co). To know more about MCoercion see #14975.
      
      We keep Refl ty as a special case for nominal reflexive coercions,
      naemly, Refl ty :: ty ~n ty.
      
      This commit is meant to be a general performance improvement,
      but there are a few regressions. See #15192, comment:13 for
      more information.
      
      Test Plan: ./validate
      
      Reviewers: bgamari, goldfire, simonpj
      
      Subscribers: rwbarton, thomie, carter
      
      GHC Trac Issues: #15192
      
      Differential Revision: https://phabricator.haskell.org/D4747
      55a3f855
  21. 19 Jun, 2018 1 commit
  22. 10 Jun, 2018 1 commit
  23. 28 May, 2018 1 commit
    • Tamar Christina's avatar
      Clean up Windows testsuite failures · 60fb2b21
      Tamar Christina authored
      Summary:
      Another round and attempt at getting these down to 0.
      
      We really should re-enable the CI and not wait for those cloud based ones.
      
      I've disabled the backpack tests on windows as they are too broad, they test
      as much the shell as they do the compiler.
      
      The perf tests have been too long to track down. but the numbers are horrible
      but I don't see them getting fixed so just have to accept them.
      
      T9293 has new windows specific output because a Dyn way only flag was added.
      This will of course not work on non-Dyn way builds.
      
      Test Plan: ./validate
      
      Reviewers: bgamari, hvr, simonmar
      
      Reviewed By: bgamari
      
      Subscribers: rwbarton, thomie, carter
      
      GHC Trac Issues: #15107
      
      Differential Revision: https://phabricator.haskell.org/D4668
      60fb2b21
  24. 24 May, 2018 1 commit
  25. 22 May, 2018 1 commit
    • Joachim Breitner's avatar
      Improve performance of CallArity · db6085b8
      Joachim Breitner authored
      the hot path contained a call to
      
          v `elemUnVarSet` (neighbors g v)
      
      and creating the set of neighbors just to check if `v` is inside
      accounted for half the allocations of the test case of #15164.
      
      By introducing a non-allocating function `hasLoopAt` for this we shave
      off half the allocations. This brings the total cost of Call Arity down
      to 20% of time and 23% of allocations, according to a profiled run. Not
      amazing, but still much better.
      
      Differential Revision: https://phabricator.haskell.org/D4718
      db6085b8
  26. 21 May, 2018 2 commits
  27. 17 May, 2018 1 commit
    • Simon Marlow's avatar
      Fix GHCi space leaks (#15111) · f27e4f62
      Simon Marlow authored
      Summary:
      There were a number of leaks causing previously loaded modules to be
      retained after a new `:load`.  This fixes enough leaks to get the
      tests to pass from D4658.
      
      Test Plan: See new tests in D4658
      
      Reviewers: niteria, bgamari, simonpj, erikd
      
      Subscribers: thomie, carter
      
      GHC Trac Issues: #15111
      
      Differential Revision: https://phabricator.haskell.org/D4659
      f27e4f62
  28. 15 May, 2018 3 commits
    • Ben Gamari's avatar
      testsuite: Disable T14697 on Windows · af986f9d
      Ben Gamari authored and Ben Gamari's avatar Ben Gamari committed
      Test Plan: Validate on Windows
      
      Subscribers: thomie, carter
      
      GHC Trac Issues: #14697, #15072
      
      Differential Revision: https://phabricator.haskell.org/D4619
      af986f9d
    • Ben Gamari's avatar
      Revert "Simplify callSiteInline a little" · 9dbf66d7
      Ben Gamari authored
      This lead to some rather significant performance regressions.
      Specifically,
      
          bytes allocated value is too high:
              Expected    T5631(normal) bytes allocated: 1106015512 +/-5%
              Lower bound T5631(normal) bytes allocated: 1050714736
              Upper bound T5631(normal) bytes allocated: 1161316288
              Actual      T5631(normal) bytes allocated: 1164953208
              Deviation   T5631(normal) bytes allocated:        5.3 %
          *** unexpected stat test failure for T5631(normal)
          max_bytes_used value is too high:
              Expected    T9630(normal) max_bytes_used: 35324712 +/-15%
              Lower bound T9630(normal) max_bytes_used: 30026005
              Upper bound T9630(normal) max_bytes_used: 40623419
              Actual      T9630(normal) max_bytes_used: 43490984
              Deviation   T9630(normal) max_bytes_used:     23.1 %
          *** unexpected stat test failure for T9630(normal)
      
      This reverts commit 7271db46.
      This reverts commit b750dcc5.
      This reverts commit 33de71fa.
      9dbf66d7
    • Ben Gamari's avatar
      testsuite: Fix expected allocations of T9020 and T12425 · df6670e7
      Ben Gamari authored
      These were both improved by d92c7556 but the
      changes were lost due to merge silliness.
      df6670e7
  29. 14 May, 2018 1 commit
    • Tobias Dammers's avatar
      Fix performance regressions from #14737 · d92c7556
      Tobias Dammers authored and Ben Gamari's avatar Ben Gamari committed
      See #15019. When removing an unnecessary type equality check in #14737,
      several regression tests failed. The cause was that some coercions that
      are actually Refl coercions weren't passed in as such, which made the
      equality check needlessly complex (Refl coercions can be discarded in
      this particular check immediately, without inspecting the types at all).
      
      We fix that, and get additional performance improvements for free.
      
      Reviewers: goldfire, bgamari, simonpj
      
      Reviewed By: bgamari, simonpj
      
      Subscribers: simonpj, thomie, carter
      
      Differential Revision: https://phabricator.haskell.org/D4635
      d92c7556
  30. 13 May, 2018 1 commit
    • Sylvain Henry's avatar
      Refactor LitString · 7c665f9c
      Sylvain Henry authored and Ben Gamari's avatar Ben Gamari committed
      Refactor LitString so that the string length is computed at most once
      and then stored.
      
      Also remove strlen and memcmp wrappers (it seems like they were a
      workaround for a very old GCC when using -fvia-C).
      
      Bumps haddock submodule.
      
      Reviewers: bgamari, dfeuer, nickkuk
      
      Reviewed By: bgamari, nickkuk
      
      Subscribers: nickkuk, dfeuer, thomie, carter
      
      Differential Revision: https://phabricator.haskell.org/D4630
      7c665f9c
  31. 06 May, 2018 1 commit
  32. 05 May, 2018 1 commit
  33. 01 May, 2018 1 commit
  34. 20 Apr, 2018 1 commit