#9281: Rewrite `integer-gmp` to use only non-allocating GMP functions --------------------------------------+------------------------------------ Reporter: hvr | Owner: hvr Type: task | Status: new Priority: normal | Milestone: 7.10.1 Component: libraries (other) | Version: Resolution: | Keywords: integer-gmp Operating System: Unknown/Multiple | Architecture: Unknown/Multiple Type of failure: None/Unknown | Difficulty: Unknown Test Case: | Blocked By: Blocking: | Related Tickets: --------------------------------------+------------------------------------ Comment (by hvr): As a brain-dump, here's the new representation I'm working on (which happens to make it trivial to provide a natural number type `Nat`): {{{#!hs type GmpLimb = Word -- actually, 'CULong' type GmpSize = Int -- actually, 'CLong' -- | Type representing /pure/ BigNats -- -- @ByteArray#@ is interpreted as an unboxed array of 'GmpLimb's -- -- Invariants: -- -- - ByteArray#'s size is the exact multiple of non-zero limbs (but never empty) -- - @0@ is represented as a 1-limb (available as 'bnZero') data BigNat = BN# ByteArray# -- | Invariant: 'NatJ#' is used iff value doesn't fit in 'NatS#' data Nat = NatS# {-# UNPACK #-} !GmpLimb | NatJ# {-# UNPACK #-} !BigNat -- | Invariant: 'Jn#'/'Jp#' are used iff value doesn't fit in 'S#' -- -- NB: less than 4 constructors allows pointer-tagging to work data Integer = S# {-# UNPACK #-} !Int | Jn# {-# UNPACK #-} !BigNat -- negative | Jp# {-# UNPACK #-} !BigNat -- positive }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9281#comment:1 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9281: Rewrite `integer-gmp` to use only non-allocating GMP functions --------------------------------------+------------------------------------ Reporter: hvr | Owner: hvr Type: task | Status: new Priority: normal | Milestone: 7.10.1 Component: libraries (other) | Version: Resolution: | Keywords: integer-gmp Operating System: Unknown/Multiple | Architecture: Unknown/Multiple Type of failure: None/Unknown | Difficulty: Unknown Test Case: | Blocked By: Blocking: | Related Tickets: --------------------------------------+------------------------------------ Comment (by simonpj): Sounds fantastic to me. Do please document the design so that others do not later accidentally slip back into calling allocating functions! Simon -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9281#comment:3 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9281: Rewrite `integer-gmp` to use only non-allocating GMP functions -------------------------------------+------------------------------------- Reporter: hvr | Owner: hvr Type: task | Status: patch Priority: normal | Milestone: 7.10.1 Component: libraries | Version: (other) | Keywords: integer-gmp Resolution: | Operating System: Unknown/Multiple Differential Revisions: Phab:D82 | Type of failure: None/Unknown Architecture: | Test Case: Unknown/Multiple | Blocking: Difficulty: Unknown | Blocked By: | Related Tickets: #8647 | -------------------------------------+------------------------------------- Comment (by simonpj): Terrific. It would be good to articulate the goal more explicitly. Is it primarily to improve interop in some way? How does it improve interop? Or is it to do with performance -- unsafe foreign calls are vastly more efficient than safe ones? Why do `integer_gmp_mpn_tdiv_q` etc need to be IO-ish at all? Can't they be pure functions? Simon -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9281#comment:5 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9281: Rewrite `integer-gmp` to use only non-allocating GMP functions -------------------------------------+------------------------------------- Reporter: hvr | Owner: hvr Type: task | Status: patch Priority: normal | Milestone: 7.10.1 Component: libraries | Version: (other) | Keywords: integer-gmp Resolution: | Operating System: Unknown/Multiple Differential Revisions: Phab:D82 | Type of failure: None/Unknown Architecture: | Test Case: Unknown/Multiple | Blocking: Difficulty: Unknown | Blocked By: | Related Tickets: #8647 | -------------------------------------+------------------------------------- Comment (by simonpj): What happens if you try {{{ foreign import ccall unsafe "integer_gmp_mpn_tdiv_q" c_mpn_tdiv_q :: MutableByteArray# s -> ByteArray# -> GmpSize# -> ByteArray# -> GmpSize# -> State# s -> State# s }}} with `-XUnliftedFFITypes`? Answer {{{ Unacceptable argument type in foreign declaration: State# s }}} But why is `State# s` unacceptable? I can think of no good reason. So I think that it might be as simple as * Adding `State#` to the list of acceptable FFI types (with `-XUnliftedFFITypes`. * Making sure that, since it's a zero-width value, we don't actually pass it to the C function. The latter step is in the code generator itself; these zero-width values ''do'' (and must) survive into `Cmm`. If you'd like to tackle this I'm sure Simon and I would help with any bits you don't grok. I think this would be MUCH simpler than trying to allow types like `IO Int#`, which has global implications that I don't know how to deal with. Simon -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9281#comment:7 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9281: Rewrite `integer-gmp` to use only non-allocating GMP functions -------------------------------------+------------------------------------- Reporter: hvr | Owner: hvr Type: task | Status: patch Priority: normal | Milestone: 7.10.1 Component: libraries | Version: (other) | Keywords: integer-gmp Resolution: | Operating System: Unknown/Multiple Differential Revisions: Phab:D82 | Type of failure: None/Unknown Architecture: | Test Case: Unknown/Multiple | Blocking: Difficulty: Unknown | Blocked By: | Related Tickets: #8647 | -------------------------------------+------------------------------------- Comment (by simonmar): `foreign import prim` has a different calling convention from `foreign import ccall` (obviously) and they look different at the `Cmm` level. The code generation for these two is quite different - the former is a `StgPrimCallOp` while the latter is a `StgFCallOp`. But I think all the machinery to do this already exists, as Simon said. When we foreign-import something with an IO type there's a wrapper around the primitive foreign call to add the IO, but the primitive call itself will have a type of the form `State# s -> (# State# s, a #)`, so we can already codegen these properly. All we need to do is make sure that there is no wrapper, which might already happen automatically if the typechecker is changed to allow `State#` arguments and `State#` or `(# State# s, a #)` results with `UnliftedFFITypes`. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9281#comment:9 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9281: Rewrite `integer-gmp` to use only non-allocating GMP functions -------------------------------------+------------------------------------- Reporter: hvr | Owner: hvr Type: task | Status: patch Priority: normal | Milestone: 7.10.1 Component: libraries | Version: (other) | Keywords: integer-gmp Resolution: | Architecture: Unknown/Multiple Operating System: | Difficulty: Unknown Unknown/Multiple | Blocked By: Type of failure: | Related Tickets: #8647 None/Unknown | Test Case: | Blocking: | Differential Revisions: Phab:D82 | -------------------------------------+------------------------------------- Comment (by Herbert Valerio Riedel ): In [changeset:"96d04186e00fe2202b9953688492bb370d818bf1/ghc"]: {{{ #!CommitTicketReference repository="ghc" revision="96d04186e00fe2202b9953688492bb370d818bf1" Remove obsolete `digitsTyConKey :: Unique` This became dead with 1e87c0a6e485e1dbef8e9ed19191e54f6cdc54e0 and was probably just missed. I plan to re-use the freed up `mkPreludeTyConUnique 23` slot soon for a new `bigNatTyConKey` (as part of the #9281 effort) }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9281#comment:11 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9281: Rewrite `integer-gmp` to use only non-allocating GMP functions -------------------------------------+------------------------------------- Reporter: hvr | Owner: hvr Type: task | Status: patch Priority: normal | Milestone: 7.10.1 Component: Core | Version: Libraries | Keywords: integer-gmp Resolution: | Architecture: Unknown/Multiple Operating System: | Difficulty: Unknown Unknown/Multiple | Blocked By: Type of failure: | Related Tickets: #8647 None/Unknown | Test Case: | Blocking: | Differential Revisions: Phab:D82 | -------------------------------------+------------------------------------- Changes (by hvr): * cc: core-libraries-committee@… (added) Old description:

After some promising results with a proof-of-concept implementation, I'm optimistic we can rewrite `integer-gmp` to use only non-allocating GMP lib functions without suffering from serious regressions.

If successful, this would

- allow to avoid the custom GMP allocator hack, and thus - avoid issues when linking against other C libraries using GMP, - simplify code, as we would perform all heap allocations in Haskell code (and never inside Cmm/C code as its done now), - and finally maybe even remove a few more superfluous temporary heap allocations.

This rewrite can be done incrementally.

New description: After some promising results with a proof-of-concept implementation, I'm optimistic we can rewrite `integer-gmp` to use only non-allocating GMP lib functions without suffering from serious regressions. If successful, this would - allow to avoid the custom GMP allocator hack, and thus - avoid issues when linking against other C libraries using GMP, - simplify code, as we would perform all heap allocations in Haskell code (and never inside Cmm/C code as its done now), - and finally maybe even remove a few more superfluous temporary heap allocations. see also wiki:Design/IntegerGmp2 -- Comment: started a wiki-page over at wiki:Design/IntegerGmp2 -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9281#comment:14 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9281: Rewrite `integer-gmp` to use only non-allocating GMP functions -------------------------------------+------------------------------------- Reporter: hvr | Owner: hvr Type: task | Status: patch Priority: normal | Milestone: 7.10.1 Component: Core | Version: Libraries | Keywords: integer-gmp Resolution: | Architecture: Unknown/Multiple Operating System: | Difficulty: Unknown Unknown/Multiple | Blocked By: Type of failure: | Related Tickets: #8647 None/Unknown | Test Case: | Blocking: | Differential Revisions: Phab:D82 | -------------------------------------+------------------------------------- Comment (by Herbert Valerio Riedel ): In [changeset:"c774b28f76ee4c220f7c1c9fd81585e0e3af0e8a/ghc"]: {{{ #!CommitTicketReference repository="ghc" revision="c774b28f76ee4c220f7c1c9fd81585e0e3af0e8a" Implement new integer-gmp2 from scratch (re #9281) This is done as a separate `integer-gmp2` backend library because it turned out to become a complete rewrite from scratch. Due to the different (over)allocation scheme and potentially different accounting (via the new `{shrink,resize}MutableByteArray#` primitives), some of the nofib benchmarks actually results in increased allocation numbers (but not necessarily an increase in runtime!). I believe the allocation numbers could improve if `{resize,shrink}MutableByteArray#` could be optimised to reallocate in-place more efficiently. Here are the more apparent changes in the latest nofib comparision between `integer-gmp` and `integer-gmp2`: ------------------------------------------------------------------ Program Size Allocs Runtime Elapsed TotalMem ------------------------------------------------------------------ ... bernouilli +1.6% +15.3% 0.132 0.132 0.0% ... cryptarithm1 -2.2% 0.0% -9.7% -9.7% 0.0% ... fasta -0.7% -0.0% +10.9% +10.9% 0.0% ... kahan +0.6% +38.9% 0.169 0.169 0.0% ... lcss -0.7% -0.0% -6.4% -6.4% 0.0% ... mandel +1.6% +33.6% 0.049 0.049 0.0% ... pidigits +0.8% +8.5% +3.9% +3.9% 0.0% power +1.4% -23.8% -18.6% -18.6% -16.7% ... primetest +1.3% +50.1% 0.085 0.085 0.0% ... rsa +1.6% +53.4% 0.026 0.026 0.0% ... scs +1.2% +6.6% +6.5% +6.6% +14.3% ... symalg +1.0% +9.5% 0.010 0.010 0.0% ... transform -0.6% -0.0% -5.9% -5.9% 0.0% ... ------------------------------------------------------------------ Min -2.3% -23.8% -18.6% -18.6% -16.7% Max +1.6% +53.4% +10.9% +10.9% +14.3% Geometric Mean -0.3% +1.9% -0.8% -0.8% +0.0% (see P35 / https://phabricator.haskell.org/P35 for full report) By default, `INTEGER_LIBRARY=integer-gmp2` is active now, which results in the package `integer-gmp-1.0.0.0` being registered in the package db. The previous `integer-gmp-0.5.1.0` can be restored by setting `INTEGER_LIBRARY=integer-gmp` (but will probably be removed altogether for GHC 7.12). In-tree GMP support has been stolen from the old `integer-gmp` (while unpatching the custom memory-allocators, as well as forcing `-fPIC`) A minor hack to `ghc-cabal` was necessary in order to support two different `integer-gmp` packages (in different folders) with the same package key. There will be a couple of follow-up commits re-implementing some features that were dropped to keep D82 minimal, as well as further clean-ups/improvements. More information can be found via #9281 and https://ghc.haskell.org/trac/ghc/wiki/Design/IntegerGmp2 Reviewed By: austin, rwbarton, simonmar Differential Revision: https://phabricator.haskell.org/D82 }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9281#comment:16 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9281: Rewrite `integer-gmp` to use only non-allocating GMP functions -------------------------------------+------------------------------------- Reporter: hvr | Owner: hvr Type: task | Status: patch Priority: normal | Milestone: 7.10.1 Component: Core | Version: Libraries | Keywords: integer-gmp Resolution: | Architecture: Unknown/Multiple Operating System: | Difficulty: Unknown Unknown/Multiple | Blocked By: Type of failure: | Related Tickets: #8647 None/Unknown | Test Case: | Blocking: | Differential Revisions: Phab:D82 | -------------------------------------+------------------------------------- Comment (by Herbert Valerio Riedel ): In [changeset:"42244668af6d8c1dd6a2d64af90ed57d8ecd8d88/ghc"]: {{{ #!CommitTicketReference repository="ghc" revision="42244668af6d8c1dd6a2d64af90ed57d8ecd8d88" Reimplement im/export primitives for integer-gmp2 The import/export operations were available in `integer-gmp-0.5.1` already, but need to be reimplemented from scratch for the `integer-gmp-1.0.0` rewrite. This also adds a few more operations than were previously available for use w/ the `BigNat` type (which will be useful for implementing serialisation for the upcoming `Natural` type) Specifically, the following operations are (re)added (albeit with slightly different type-signatures): - `sizeInBaseBigNat` - `sizeInBaseInteger` - `sizeInBaseWord#` - `exportBigNatToAddr` - `exportIntegerToAddr` - `exportWordToAddr` - `exportBigNatToMutableByteArray` - `exportIntegerToMutableByteArray` - `exportWordToMutableByteArray` - `importBigNatFromAddr` - `importIntegerFromAddr` - `importBigNatFromByteArray` - `importIntegerFromByteArray` NOTE: The `integerGmpInternals` test-case is updated but not yet re-enabled as it contains tests for other primitives which aren't yet reimplemented. This addresses #9281 Reviewed By: austin, duncan Differential Revision: https://phabricator.haskell.org/D480 }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9281#comment:18 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9281: Rewrite `integer-gmp` to use only non-allocating GMP functions -------------------------------------+------------------------------------- Reporter: hvr | Owner: hvr Type: task | Status: patch Priority: normal | Milestone: 7.10.1 Component: Core | Version: Libraries | Keywords: integer-gmp Resolution: | Architecture: Unknown/Multiple Operating System: | Difficulty: Unknown Unknown/Multiple | Blocked By: Type of failure: | Related Tickets: #8647 None/Unknown | Test Case: | Blocking: | Differential Revisions: Phab:D82 | -------------------------------------+------------------------------------- Comment (by Herbert Valerio Riedel ): In [changeset:"be7fb7e58c70cd9b0a933fb26cd5f2607d6dc4b2/ghc"]: {{{ #!CommitTicketReference repository="ghc" revision="be7fb7e58c70cd9b0a933fb26cd5f2607d6dc4b2" integer-gmp2: export `Word`-counterpart of gcdInt It's trivial for `integer-gmp2` (#9281) to provide it, and it'll be useful for a future 'Natural'-related commit, as well as providing a `Word` optimised `gcd`-RULE. }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9281#comment:19 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9281: Rewrite `integer-gmp` to use only non-allocating GMP functions -------------------------------------+------------------------------------- Reporter: hvr | Owner: hvr Type: task | Status: patch Priority: normal | Milestone: 7.10.1 Component: Core | Version: Libraries | Keywords: integer-gmp Resolution: | Architecture: Unknown/Multiple Operating System: | Difficulty: Unknown Unknown/Multiple | Blocked By: Type of failure: | Related Tickets: #8647 None/Unknown | Test Case: | Blocking: | Differential Revisions: Phab:D82 | -------------------------------------+------------------------------------- Comment (by Herbert Valerio Riedel ): In [changeset:"a9a0dd34dcdfb7309f57bda88435acca14ec54d5/ghc"]: {{{ #!CommitTicketReference repository="ghc" revision="a9a0dd34dcdfb7309f57bda88435acca14ec54d5" Install `ghc-gmp.h` C include header file (#9281) This is mostly interesting when using the in-tree GMP, as there's no way otherwise to access the in-tree `gmp.h` header file after installation. In case `integer-gmp2` was build against a system-installed GMP library, `ghc-gmp.h` simply contains `#include ` for convenience. Reviewed By: ekmett Differential Revision: https://phabricator.haskell.org/D522 }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9281#comment:21 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9281: Rewrite `integer-gmp` to use only non-allocating GMP functions -------------------------------------+------------------------------------- Reporter: hvr | Owner: hvr Type: task | Status: patch Priority: normal | Milestone: 7.10.1 Component: Core | Version: Libraries | Keywords: integer-gmp Resolution: | Architecture: Unknown/Multiple Operating System: | Difficulty: Unknown Unknown/Multiple | Blocked By: Type of failure: | Related Tickets: #8647 None/Unknown | Test Case: | Blocking: | Differential Revisions: Phab:D82 | -------------------------------------+------------------------------------- Comment (by Herbert Valerio Riedel ): In [changeset:"58dcd5c2e2a94643454296ea0bb109db96bd154f/ghc"]: {{{ #!CommitTicketReference repository="ghc" revision="58dcd5c2e2a94643454296ea0bb109db96bd154f" Re-implement `testPrimeInteger` predicate (#9281) This also adds `testPrimeWord#` and `testPrimeBigNat` predicates. `testPrimeInteger` has been available since `integer-gmp-0.5.1` (added via f49735486533842cc84df70cafc8d565dffd75db). The `nextPrimeInteger` function is still missing though. }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9281#comment:24 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9281: Rewrite `integer-gmp` to use only non-allocating GMP functions -------------------------------------+------------------------------------- Reporter: hvr | Owner: hvr Type: task | Status: patch Priority: normal | Milestone: 7.10.1 Component: Core | Version: Libraries | Keywords: integer-gmp Resolution: | Architecture: Unknown/Multiple Operating System: | Difficulty: Unknown Unknown/Multiple | Blocked By: Type of failure: | Related Tickets: #8647 None/Unknown | Test Case: | Blocking: | Differential Revisions: Phab:D82 | -------------------------------------+------------------------------------- Comment (by Herbert Valerio Riedel ): In [changeset:"8d7831108644584f710515b39b9fc97fbeca7a4c/ghc"]: {{{ #!CommitTicketReference repository="ghc" revision="8d7831108644584f710515b39b9fc97fbeca7a4c" Re-implement `nextPrimeInteger` predicate (#9281) This also adds `nextPrimeWord#` and `nextPrimeBigNat` predicates. `nextPrimeInteger` has been available since `integer-gmp-0.5.1` (added via f49735486533842cc84df70cafc8d565dffd75db). }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9281#comment:25 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9281: Rewrite `integer-gmp` to use only non-allocating GMP functions -------------------------------------+------------------------------------- Reporter: hvr | Owner: hvr Type: task | Status: patch Priority: normal | Milestone: 7.10.1 Component: Core | Version: Libraries | Keywords: integer-gmp Resolution: | Architecture: Unknown/Multiple Operating System: | Difficulty: Unknown Unknown/Multiple | Blocked By: Type of failure: | Related Tickets: #8647 None/Unknown | Test Case: | Blocking: | Differential Revisions: Phab:D82 | -------------------------------------+------------------------------------- Comment (by Herbert Valerio Riedel ): In [changeset:"83c48438c800986c537d3cae682d53ee8ed326ed/ghc"]: {{{ #!CommitTicketReference repository="ghc" revision="83c48438c800986c537d3cae682d53ee8ed326ed" Re-implement `recipModInteger` (#9281) This also exposes the following two type-specialised modular exponentiation variants of `recipModInteger` useful for implementing a `recipModNatural` operation. recipModBigNat :: BigNat -> BigNat -> BigNat recipModWord :: Word# -> Word# -> Word# `recipModInteger` has been available since `integer-gmp-0.5.1` (added via 4d516855241b70eb687d95e3c121428de885e83e) }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9281#comment:28 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9281: Rewrite `integer-gmp` to use only non-allocating GMP functions -------------------------------------+------------------------------------- Reporter: hvr | Owner: hvr Type: task | Status: closed Priority: normal | Milestone: 7.10.1 Component: Core | Version: Libraries | Keywords: integer-gmp Resolution: fixed | Architecture: Unknown/Multiple Operating System: | Difficulty: Unknown Unknown/Multiple | Blocked By: Type of failure: | Related Tickets: #8647 None/Unknown | Test Case: | Blocking: | Differential Revisions: Phab:D82 | -------------------------------------+------------------------------------- Changes (by hvr): * status: patch => closed * resolution: => fixed Comment: Let's consider this accomplished for now. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9281#comment:30 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#9281: Rewrite `integer-gmp` to use only non-allocating GMP functions -------------------------------------+------------------------------------- Reporter: hvr | Owner: hvr Type: task | Status: closed Priority: normal | Milestone: 7.10.1 Component: Core Libraries | Version: Resolution: fixed | Keywords: integer- Operating System: Unknown/Multiple | gmp Type of failure: None/Unknown | Architecture: Blocked By: | Unknown/Multiple Related Tickets: #8647 #9818 | Test Case: | Blocking: | Differential Revisions: Phab:D82 -------------------------------------+------------------------------------- Comment (by Herbert Valerio Riedel ): In [changeset:"995e8c1c8692b60c907c7d2ccea179d52ca8e69e/ghc"]: {{{ #!CommitTicketReference repository="ghc" revision="995e8c1c8692b60c907c7d2ccea179d52ca8e69e" Drop old integer-gmp-0.5 from GHC source tree This completes what c774b28f76ee4c220f7c1c9fd81585e0e3af0e8a (#9281) started. `integer-gmp-1.0` was added as an additional `libraries/integer-gmp2` folder while retaining the ability to configure GHC w/ the old `integer-gmp-0.5` to have a way back, and or the ability to easily switch between old/new `integer-gmp` for benchmark/debugging purposes. This commit removes the old `libraries/integer-gmp` folder and moves `libraries/integer-gmp2` into its place, while removing any mentions of "gmp2" as well as the to support two different `integer-gmp` packages in GHC's source-tree. Reviewed By: austin Differential Revision: https://phabricator.haskell.org/D769 }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9281#comment:32 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler