[Git][ghc/ghc][wip/nested-cpr-2019] 81 commits: Remove duplicate Note [When to print foralls] in GHC.Core.TyCo.Ppr

Sebastian Graf gitlab at gitlab.haskell.org
Tue May 26 09:26:58 UTC 2020 


Sebastian Graf pushed to branch wip/nested-cpr-2019 at Glasgow Haskell Compiler / GHC


Commits:
5bcf8606 by Ryan Scott at 2020-05-17T08:46:38-04:00
Remove duplicate Note [When to print foralls] in GHC.Core.TyCo.Ppr

There are two different Notes named `[When to print foralls]`. The
most up-to-date one is in `GHC.Iface.Type`, but there is a second
one in `GHC.Core.TyCo.Ppr`. The latter is less up-to-date, as it was
written before GHC switched over to using ifaces to pretty-print
types. I decided to just remove the latter and replace it with a
reference to the former.

[ci skip]

- - - - -
55f0e783 by Fumiaki Kinoshita at 2020-05-21T12:10:44-04:00
base: Add Generic instances to various datatypes under GHC.*

* GHC.Fingerprint.Types: Fingerprint
* GHC.RTS.Flags: GiveGCStats, GCFlags, ConcFlags, DebugFlags, CCFlags, DoHeapProfile, ProfFlags, DoTrace, TraceFlags, TickyFlags, ParFlags and RTSFlags
* GHC.Stats: RTSStats and GCStats
* GHC.ByteOrder: ByteOrder
* GHC.Unicode: GeneralCategory
* GHC.Stack.Types: SrcLoc

Metric Increase:
    haddock.base

- - - - -
a9311cd5 by Gert-Jan Bottu at 2020-05-21T12:11:31-04:00
Explicit Specificity

Implementation for Ticket #16393.
Explicit specificity allows users to manually create inferred type variables,
by marking them with braces.
This way, the user determines which variables can be instantiated through
visible type application.

The additional syntax is included in the parser, allowing users to write
braces in type variable binders (type signatures, data constructors etc).
This information is passed along through the renamer and verified in the
type checker.
The AST for type variable binders, data constructors, pattern synonyms,
partial signatures and Template Haskell has been updated to include the
specificity of type variables.

Minor notes:
- Bumps haddock submodule
- Disables pattern match checking in GHC.Iface.Type with GHC 8.8

- - - - -
24e61aad by Ben Price at 2020-05-21T12:12:17-04:00
Lint should say when it is checking a rule

It is rather confusing that when lint finds an error in a rule attached
to a binder, it reports the error as in the RHS, not the rule:
  ...
  In the RHS of foo

We add a clarifying line:
  ...
  In the RHS of foo
  In a rule attached to foo

The implication that the rule lives inside the RHS is a bit odd, but
this niggle is already present for unfoldings, whose pattern we are
following.

- - - - -
78c6523c by Ben Gamari at 2020-05-21T12:13:01-04:00
nonmoving: Optimise the write barrier

- - - - -
13f6c9d0 by Andreas Klebinger at 2020-05-21T12:13:45-04:00
Refactor linear reg alloc to remember past assignments.

When assigning registers we now first try registers we
assigned to in the past, instead of picking the "first"
one.

This is in extremely helpful when dealing with loops for
which variables are dead for part of the loop.

This is important for patterns like this:

        foo = arg1
    loop:
        use(foo)
        ...
        foo = getVal()
        goto loop;

There we:
* assign foo to the register of arg1.
* use foo, it's dead after this use as it's overwritten after.
* do other things.
* look for a register to put foo in.

If we pick an arbitrary one it might differ from the register the
start of the loop expect's foo to be in.
To fix this we simply look for past register assignments for
the given variable. If we find one and the register is free we
use that register.

This reduces the need for fixup blocks which match the register
assignment between blocks. In the example above between the end
and the head of the loop.

This patch also moves branch weight estimation ahead of register
allocation and adds a flag to control it (cmm-static-pred).
* It means the linear allocator is more likely to assign the hotter
  code paths first.
* If it assign these first we are:
  + Less likely to spill on the hot path.
  + Less likely to introduce fixup blocks on the hot path.

These two measure combined are surprisingly effective. Based on nofib
we get in the mean:

* -0.9% instructions executed
* -0.1% reads/writes
* -0.2% code size.
* -0.1% compiler allocations.
* -0.9% compile time.
* -0.8% runtime.

Most of the benefits are simply a result of removing redundant moves
and spills.

Reduced compiler allocations likely are the result of less code being
generated. (The added lookup is mostly non-allocating).

- - - - -
edc2cc58 by Andreas Klebinger at 2020-05-21T12:14:25-04:00
NCG: Codelayout: Distinguish conditional and other branches.

In #18053 we ended up with a suboptimal code layout because
the code layout algorithm didn't distinguish between conditional
and unconditional control flow.

We can completely eliminate unconditional control flow instructions
by placing blocks next to each other, not so much for conditionals.

In terms of implementation we simply give conditional branches less
weight before computing the layout.

Fixes #18053

- - - - -
b7a6b2f4 by Gleb Popov at 2020-05-21T12:15:26-04:00
gitlab-ci: Set locale to C.UTF-8.

- - - - -
a8c27cf6 by Stefan Holdermans at 2020-05-21T12:16:08-04:00
Allow spaces in GHCi :script file names

This patch updates the user interface of GHCi so that file names passed
to the ':script' command may contain spaces escaped with a backslash.

For example:

  :script foo\ bar.script

The implementation uses a modified version of 'words' that does not
break on escaped spaces.

Fixes #18027.

- - - - -
82663959 by Stefan Holdermans at 2020-05-21T12:16:08-04:00
Add extra tests for GHCi :script syntax checks

The syntax for GHCi's ":script" command allows for only a single file
name to be passed as an argument. This patch adds a test for the cases
in which a file name is missing or multiple file names are passed.

Related to #T18027.

- - - - -
a0b79e1b by Stefan Holdermans at 2020-05-21T12:16:08-04:00
Allow GHCi :script file names in double quotes

This patch updates the user interface of GHCi so that file names passed
to the ':script' command can be wrapped in double quotes.

For example:

  :script "foo bar.script"

The implementation uses a modified version of 'words' that treats
character sequences enclosed in double quotes as single words.

Fixes #18027.

- - - - -
cf566330 by Stefan Holdermans at 2020-05-21T12:16:08-04:00
Update documentation for GHCi :script

This patch adds the fixes that allow for file names containing spaces to
be passed to GHCi's ':script' command to the release notes for 8.12 and
expands the user-guide documentation for ':script' by mentioning how
such file names can be passed.

Related to #18027.

- - - - -
0004ccb8 by Tuan Le at 2020-05-21T12:16:46-04:00
llvmGen: Consider Relocatable read-only data as not constantReferences: #18137

- - - - -
964d3ea2 by John Ericson at 2020-05-21T12:17:30-04:00
Use `Checker` for `tc_pat`

- - - - -
b797aa42 by John Ericson at 2020-05-21T12:17:30-04:00
Use `Checker` for `tc_lpat` and `tc_lpats`

- - - - -
5108e84a by John Ericson at 2020-05-21T12:17:30-04:00
More judiciously panic in `ts_pat`

- - - - -
510e0451 by John Ericson at 2020-05-21T12:17:30-04:00
Put `PatEnv` first in `GHC.Tc.Gen.Pat.Checker`

- - - - -
cb4231db by John Ericson at 2020-05-21T12:17:30-04:00
Tiny cleaup eta-reduce away a function argument

In GHC, not in the code being compiled!

- - - - -
6890c38d by John Ericson at 2020-05-21T12:17:30-04:00
Use braces with do in `SplicePat` case for consistency

- - - - -
3451584f by buggymcbugfix at 2020-05-21T12:18:06-04:00
Fix spelling mistakes and typos

- - - - -
b552e531 by buggymcbugfix at 2020-05-21T12:18:06-04:00
Add INLINABLE pragmas to Enum list producers

The INLINABLE pragmas ensure that we export stable (unoptimised) unfoldings in
the interface file so we can do list fusion at usage sites.

Related tickets: #15185, #8763, #18178.

- - - - -
e7480063 by buggymcbugfix at 2020-05-21T12:18:06-04:00
Piggyback on Enum Word methods for Word64

If we are on a 64 bit platform, we can use the efficient Enum Word
methods for the Enum Word64 instance.

- - - - -
892b0c41 by buggymcbugfix at 2020-05-21T12:18:06-04:00
Document INLINE(ABLE) pragmas that enable fusion

- - - - -
2b363ebb by Richard Eisenberg at 2020-05-21T12:18:45-04:00
MR template should ask for key part
- - - - -
a95bbd0b by Sebastian Graf at 2020-05-21T12:19:37-04:00
Make `Int`'s `mod` and `rem` strict in their first arguments

They used to be strict until 4d2ac2d (9 years ago).

It's obviously better to be strict for performance reasons.
It also blocks #18067.

NoFib results:

```
--------------------------------------------------------------------------------
        Program         Allocs    Instrs
--------------------------------------------------------------------------------
        integer          -1.1%     +0.4%
   wheel-sieve2         +21.2%    +20.7%
--------------------------------------------------------------------------------
            Min          -1.1%     -0.0%
            Max         +21.2%    +20.7%
 Geometric Mean          +0.2%     +0.2%
```

The regression in `wheel-sieve2` is due to reboxing that likely will go
away with the resolution of #18067. See !3282 for details.

Fixes #18187.

- - - - -
d3d055b8 by Galen Huntington at 2020-05-21T12:20:18-04:00
Clarify pitfalls of NegativeLiterals; see #18022.
- - - - -
1b508a9e by Alexey Kuleshevich at 2020-05-21T12:21:02-04:00
Fix wording in primops documentation to reflect the correct reasoning:

* Besides resizing functions, shrinking ones also mutate the
  size of a mutable array and because of those two `sizeofMutabeByteArray`
  and `sizeofSmallMutableArray` are now deprecated
* Change reference in documentation to the newer functions `getSizeof*`
  instead of `sizeof*` for shrinking functions
* Fix incorrect mention of "byte" instead of "small"

- - - - -
4ca0c8a1 by Andreas Klebinger at 2020-05-21T12:21:53-04:00
Don't variable-length encode magic iface constant.

We changed to use variable length encodings for many types by default,
including Word32. This makes sense for numbers but not when Word32 is
meant to represent four bytes.

I added a FixedLengthEncoding newtype to Binary who's instances
interpret their argument as a collection of bytes instead of a number.

We then use this when writing/reading magic numbers to the iface file.

I also took the libery to remove the dummy iface field.

This fixes #18180.

- - - - -
a1275081 by Krzysztof Gogolewski at 2020-05-21T12:22:35-04:00
Add a regression test for #11506

The testcase works now.
See explanation in https://gitlab.haskell.org/ghc/ghc/issues/11506#note_273202

- - - - -
8a816e5f by Krzysztof Gogolewski at 2020-05-21T12:23:55-04:00
Sort deterministically metric output

Previously, we sorted according to the test name and way,
but the metrics (max_bytes_used/peak_megabytes_allocated etc.)
were appearing in nondeterministic order.

- - - - -
566cc73f by Sylvain Henry at 2020-05-21T12:24:45-04:00
Move isDynLinkName into GHC.Types.Name

It doesn't belong into GHC.Unit.State

- - - - -
d830bbc9 by Adam Sandberg Ericsson at 2020-05-23T13:36:20-04:00
docs: fix formatting and add some links

[skip ci]

- - - - -
49301ad6 by Andrew Martin at 2020-05-23T13:37:01-04:00
Implement cstringLength# and FinalPtr

This function and its accompanying rule resolve issue #5218.
A future PR to the bytestring library will make the internal
Data.ByteString.Internal.unsafePackAddress compute string length
with cstringLength#. This will improve the status quo because it is
eligible for constant folding.

Additionally, introduce a new data constructor to ForeignPtrContents
named FinalPtr. This additional data constructor, when used in the
IsString instance for ByteString, leads to more Core-to-Core
optimization opportunities, fewer runtime allocations, and smaller
binaries.

Also, this commit re-exports all the functions from GHC.CString
(including cstringLength#) in GHC.Exts. It also adds a new test
driver. This test driver is used to perform substring matches on Core
that is dumped after all the simplifier passes. In this commit, it is
used to check that constant folding of cstringLength# works.

- - - - -
dcd6bdcc by Ben Gamari at 2020-05-23T13:37:48-04:00
simplCore: Ignore ticks in rule templates

This fixes #17619, where a tick snuck in to the template of a rule,
resulting in a panic during rule matching. The tick in question was
introduced via post-inlining, as discussed in `Note [Simplifying
rules]`. The solution we decided upon was to simply ignore ticks in the
rule template, as discussed in `Note [Tick annotations in RULE
matching]`.

Fixes #18162.
Fixes #17619.

- - - - -
82cb8913 by John Ericson at 2020-05-23T13:38:32-04:00
Fix #18145 and also avoid needless work with implicit vars

 - `forAllOrNothing` now is monadic, so we can trace whether we bind
   an explicit `forall` or not.

 - #18145 arose because the free vars calculation was needlessly
   complex. It is now greatly simplified.

 - Replaced some other implicit var code with `filterFreeVarsToBind`.

Co-authored-by: Ryan Scott <ryan.gl.scott at gmail.com>

- - - - -
a60dc835 by Ben Gamari at 2020-05-23T13:39:12-04:00
Bump process submodule

Fixes #17926.

- - - - -
856adf54 by Ben Gamari at 2020-05-23T13:40:21-04:00
users-guide: Clarify meaning of -haddock flag

Fixes #18206.

- - - - -
7ae57afd by Ben Gamari at 2020-05-23T13:41:03-04:00
git: Add ignored commits file

This can be used to tell git to ignore bulk renaming commits like the
recently-finished module hierarchy refactoring. Configured with,

    git config blame.ignoreRevsFile .git-ignore-revs

- - - - -
63d30e60 by jneira at 2020-05-24T01:54:42-04:00
Add hie-bios script for windows systems
It is a direct translation of the sh script

- - - - -
59182b88 by jneira at 2020-05-24T01:54:42-04:00
Honour previous values for CABAL and CABFLAGS
The immediate goal is let the hie-bios.bat script
set CABFLAGS with `-v0` and remove all cabal output
except the compiler arguments

- - - - -
932dc54e by jneira at 2020-05-24T01:54:42-04:00
Add specific configuration for windows in hie.yaml

- - - - -
e0eda070 by jneira at 2020-05-24T01:54:42-04:00
Remove not needed hie-bios output

- - - - -
a0ea59d6 by Sylvain Henry at 2020-05-24T01:55:24-04:00
Move Config module into GHC.Settings

- - - - -
37430251 by Sylvain Henry at 2020-05-24T01:55:24-04:00
Rename GHC.Core.Arity into GHC.Core.Opt.Arity

- - - - -
a426abb9 by Sylvain Henry at 2020-05-24T01:55:24-04:00
Rename GHC.Hs.Types into GHC.Hs.Type

See discussion in https://gitlab.haskell.org/ghc/ghc/issues/13009#note_268610

- - - - -
1c91a7a0 by Sylvain Henry at 2020-05-24T01:55:24-04:00
Bump haddock submodule

- - - - -
66bd24d1 by Ryan Scott at 2020-05-24T01:56:03-04:00
Add orderingTyCon to wiredInTyCons (#18185)

`Ordering` needs to be wired in for use in the built-in `CmpNat` and
`CmpSymbol` type families, but somehow it was never added to the list
of `wiredInTyCons`, leading to the various oddities observed
in #18185. Easily fixed by moving `orderingTyCon` from
`basicKnownKeyNames` to `wiredInTyCons`.

Fixes #18185.

- - - - -
01c43634 by Matthew Pickering at 2020-05-24T01:56:42-04:00
Remove unused hs-boot file

- - - - -
7a07aa71 by Sylvain Henry at 2020-05-24T15:22:17-04:00
Hadrian: fix cross-compiler build (#16051)

- - - - -
15ccca16 by Sylvain Henry at 2020-05-24T15:22:17-04:00
Hadrian: fix distDir per stage

- - - - -
b420fb24 by Sylvain Henry at 2020-05-24T15:22:17-04:00
Hadrian: fix hp2ps error during cross-compilation

Fixed by @alp (see https://gitlab.haskell.org/ghc/ghc/issues/16051#note_274265)

- - - - -
cd339ef0 by Joshua Price at 2020-05-24T15:22:56-04:00
Make Unicode brackets opening/closing tokens (#18225)

The tokens `[|`, `|]`, `(|`, and `|)` are opening/closing tokens as
described in GHC Proposal #229. This commit makes the unicode
variants (`⟦`, `⟧`, `⦇`, and `⦈`) act the same as their ASCII
counterparts.

- - - - -
61a46464 by Sebastian Graf at 2020-05-26T11:26:06+02:00
Nested CPR

- - - - -
56014585 by Sebastian Graf at 2020-05-26T11:26:06+02:00
Move tests from stranal to cpranal

- - - - -
32bab299 by Sebastian Graf at 2020-05-26T11:26:06+02:00
Accept FacState

- - - - -
5b452873 by Sebastian Graf at 2020-05-26T11:26:06+02:00
Factor Cpr and Termination into a joint lattice

As a result, we don't even have to export Termination from Cpr. Neat!

Also I realised there is a simpler and more sound way to generate and
unleash CPR signatures.

- - - - -
051e28e1 by Sebastian Graf at 2020-05-26T11:26:42+02:00
Consider unboxing effects of WW better and get rid of hack

- - - - -
4d0c1f9d by Sebastian Graf at 2020-05-26T11:26:44+02:00
stuff

- - - - -
0487b064 by Sebastian Graf at 2020-05-26T11:26:44+02:00
Debug output

- - - - -
630ce44d by Sebastian Graf at 2020-05-26T11:26:44+02:00
A slew of testsuite changes

- - - - -
f1234f76 by Sebastian Graf at 2020-05-26T11:26:44+02:00
Fix T1600

- - - - -
3f3d7881 by Sebastian Graf at 2020-05-26T11:26:44+02:00
Inline `decodeDoubleInteger` and constant-fold `decodeDouble_Int64#` instead

Currently, `decodeDoubleInteger` is known-key so that it can be
recognised in constant folding. But that is very brittle and doesn't
survive worker/wrapper, which we even do for
`NOINLINE`/`CONSTANT_FOLDED` things since #13143.
Also it is a trade-off: The implementation of `decodeDoubleInteger`
allocates an `Integer` box that never cancels aways if we don't inline
it.

Hence we recognise the `decodeDouble_Int64#` primop instead in constant
folding, so that we can inline `decodeDoubleInteger`.

You may wonder how this affects performance of code using
`integer-simple`; Apparently, according to @hsyl20 this is not a concern
since we will hopefully land !2231 soon.

Fixes #18092.

- - - - -
2e4bb983 by Sebastian Graf at 2020-05-26T11:26:44+02:00
Fix primop termination

- - - - -
8dbfd893 by Sebastian Graf at 2020-05-26T11:26:44+02:00
Test for DataCon wrapper CPR

- - - - -
f9386552 by Sebastian Graf at 2020-05-26T11:26:45+02:00
Fix CPR of bottoming functions/primops

- - - - -
2472d714 by Sebastian Graf at 2020-05-26T11:26:45+02:00
Fix DataConWrapperCpr and accept other test outputs

- - - - -
e8aabb57 by Sebastian Graf at 2020-05-26T11:26:45+02:00
Accept two more changed test outputs

- - - - -
d0a9c046 by Sebastian Graf at 2020-05-26T11:26:45+02:00
Update CaseBinderCPR with a new function

- - - - -
8143ce54 by Sebastian Graf at 2020-05-26T11:26:45+02:00
Don't give the case binder the CPR property

- - - - -
0dfaf57e by Sebastian Graf at 2020-05-26T11:26:45+02:00
Prune CPR sigs to constant depth on all bindings

- - - - -
05642eb3 by Sebastian Graf at 2020-05-26T11:26:45+02:00
Use variable length coding for ConTags

- - - - -
159777d4 by Sebastian Graf at 2020-05-26T11:26:45+02:00
Accept testuite output

- - - - -
9b391fd5 by Sebastian Graf at 2020-05-26T11:26:45+02:00
Don't attach CPR sigs to expandable bindings; transform their unfoldings instead

- - - - -
32f29281 by Sebastian Graf at 2020-05-26T11:26:45+02:00
Revert "Don't give the case binder the CPR property"

This reverts commit 910edd76d5fe68b58c74f3805112f9faef4f2788.

It seems we broke too much with this change. We lost our big win in
`fish`.

- - - - -
13545e2d by Sebastian Graf at 2020-05-26T11:26:45+02:00
A more modular and configurable approach to optimistic case binder CPR

- - - - -
545ba9d1 by Sebastian Graf at 2020-05-26T11:26:45+02:00
Fix T9291

- - - - -
b58def22 by Sebastian Graf at 2020-05-26T11:26:45+02:00
Document -fcase-binder-cpr-depth in the user's guide

- - - - -
bb965505 by Sebastian Graf at 2020-05-26T11:26:45+02:00
Rebase woes

- - - - -
69fe0bd6 by Sebastian Graf at 2020-05-26T11:26:45+02:00
Testsuite changes

- - - - -
d33b8ba7 by Sebastian Graf at 2020-05-26T11:26:45+02:00
Refactoring around cprAnalBind

- - - - -
2594ade0 by Sebastian Graf at 2020-05-26T11:26:45+02:00
Fix case binder CPR by not looking into unfoldings of case binders

- - - - -


30 changed files:

- + .git-ignore-revs
- .gitlab-ci.yml
- .gitlab/ci.sh
- .gitlab/merge_request_templates/merge-request.md
- compiler/GHC/Builtin/Names.hs
- − compiler/GHC/Builtin/Names.hs-boot
- compiler/GHC/Builtin/Names/TH.hs
- compiler/GHC/Builtin/Types.hs
- compiler/GHC/Builtin/primops.txt.pp
- compiler/GHC/Cmm.hs
- compiler/GHC/Cmm/CLabel.hs
- compiler/GHC/CmmToAsm.hs
- compiler/GHC/CmmToAsm/BlockLayout.hs
- compiler/GHC/CmmToAsm/CFG.hs
- compiler/GHC/CmmToAsm/Dwarf.hs
- compiler/GHC/CmmToAsm/Instr.hs
- compiler/GHC/CmmToAsm/Reg/Linear.hs
- compiler/GHC/CmmToAsm/Reg/Linear/Base.hs
- compiler/GHC/CmmToAsm/Reg/Linear/PPC.hs
- compiler/GHC/CmmToAsm/Reg/Linear/SPARC.hs
- compiler/GHC/CmmToAsm/Reg/Linear/State.hs
- compiler/GHC/CmmToAsm/Reg/Linear/X86.hs
- compiler/GHC/CmmToAsm/Reg/Linear/X86_64.hs
- compiler/GHC/CmmToC.hs
- compiler/GHC/CmmToLlvm/Data.hs
- compiler/GHC/Core/Coercion.hs
- compiler/GHC/Core/ConLike.hs
- compiler/GHC/Core/DataCon.hs
- compiler/GHC/Core/DataCon.hs-boot
- compiler/GHC/Core/Lint.hs


The diff was not included because it is too large.


View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/3ae7eb1f530eba8e30fb85a0f5d018e15666b882...2594ade04814abc16db74367265be92acc82c94a

-- 
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/compare/3ae7eb1f530eba8e30fb85a0f5d018e15666b882...2594ade04814abc16db74367265be92acc82c94a
You're receiving this email because of your account on gitlab.haskell.org.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/ghc-commits/attachments/20200526/e6219141/attachment-0001.html>

More information about the ghc-commits mailing list