[compiler-rt] [clang] [llvm] [libcxx] [flang] [clang-tools-extra] [libc] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

@@ -29,12 +30,11 @@
 
 _LIBCPP_BEGIN_NAMESPACE_STD
 
-template <
-typename _DifferenceType,
-typename _Tp,
-typename _BinaryOperation,
-typename _UnaryOperation,
-__enable_if_t<__is_trivial_plus_operation<_BinaryOperation, _Tp, 
_Tp>::value && is_arithmetic_v<_Tp>, int> = 0>
+template < typename _DifferenceType,
+   typename _Tp,
+   typename _BinaryOperation,
+   typename _UnaryOperation,
+   __enable_if_t<__desugars_to<_BinaryOperation, plus<_Tp> >::value && 
is_arithmetic_v<_Tp>, int> = 0>

AntonRydahl wrote:

I tried to remove is_arithmetic_type, and it makes some of our tests fail.

https://github.com/llvm/llvm-project/pull/68642
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[flang] [compiler-rt] [libc] [libcxx] [clang-tools-extra] [llvm] [clang] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/68642

>From f0d93cc6a5cd485c654ab38221691db038bacc7d Mon Sep 17 00:00:00 2001
From: AntonRydahl 
Date: Mon, 9 Oct 2023 15:13:22 -0700
Subject: [PATCH 1/6] Merged __is_trivial_equality_predicate and
 __is_trivial_plus_operation into __desugars_to

---
 libcxx/include/CMakeLists.txt |  1 -
 libcxx/include/__algorithm/comp.h |  7 ++---
 libcxx/include/__algorithm/equal.h| 22 
 .../cpu_backends/transform_reduce.h   | 23 
 libcxx/include/__functional/operations.h  | 15 +--
 .../include/__functional/ranges_operations.h  |  7 ++---
 .../include/__numeric/pstl_transform_reduce.h |  2 +-
 .../include/__type_traits/operation_traits.h  |  4 +--
 .../include/__type_traits/predicate_traits.h  | 26 ---
 libcxx/include/module.modulemap.in|  1 -
 10 files changed, 40 insertions(+), 68 deletions(-)
 delete mode 100644 libcxx/include/__type_traits/predicate_traits.h

diff --git a/libcxx/include/CMakeLists.txt b/libcxx/include/CMakeLists.txt
index 340353f8ebb41c4..7eb09a06ccd482e 100644
--- a/libcxx/include/CMakeLists.txt
+++ b/libcxx/include/CMakeLists.txt
@@ -813,7 +813,6 @@ set(files
   __type_traits/negation.h
   __type_traits/noexcept_move_assign_container.h
   __type_traits/operation_traits.h
-  __type_traits/predicate_traits.h
   __type_traits/promote.h
   __type_traits/rank.h
   __type_traits/remove_all_extents.h
diff --git a/libcxx/include/__algorithm/comp.h 
b/libcxx/include/__algorithm/comp.h
index 9474536615ffb67..0993c37cce36a6b 100644
--- a/libcxx/include/__algorithm/comp.h
+++ b/libcxx/include/__algorithm/comp.h
@@ -10,8 +10,9 @@
 #define _LIBCPP___ALGORITHM_COMP_H
 
 #include <__config>
+#include <__functional/operations.h>
 #include <__type_traits/integral_constant.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
 #  pragma GCC system_header
@@ -26,8 +27,8 @@ struct __equal_to {
   }
 };
 
-template 
-struct __is_trivial_equality_predicate<__equal_to, _Lhs, _Rhs> : true_type {};
+template <>
+struct __desugars_to<__equal_to, std::equal_to<>> : true_type {};
 
 // The definition is required because __less is part of the ABI, but it's empty
 // because all comparisons should be transparent.
diff --git a/libcxx/include/__algorithm/equal.h 
b/libcxx/include/__algorithm/equal.h
index b69aeff92bb9289..35e82da15e4d058 100644
--- a/libcxx/include/__algorithm/equal.h
+++ b/libcxx/include/__algorithm/equal.h
@@ -15,6 +15,7 @@
 #include <__config>
 #include <__functional/identity.h>
 #include <__functional/invoke.h>
+#include <__functional/operations.h>
 #include <__iterator/distance.h>
 #include <__iterator/iterator_traits.h>
 #include <__string/constexpr_c_functions.h>
@@ -23,7 +24,7 @@
 #include <__type_traits/is_constant_evaluated.h>
 #include <__type_traits/is_equality_comparable.h>
 #include <__type_traits/is_volatile.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 #include <__utility/move.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
@@ -41,13 +42,12 @@ _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI 
_LIBCPP_CONSTEXPR_SINCE_CXX20 boo
   return true;
 }
 
-template <
-class _Tp,
-class _Up,
-class _BinaryPredicate,
-__enable_if_t<__is_trivial_equality_predicate<_BinaryPredicate, _Tp, 
_Up>::value && !is_volatile<_Tp>::value &&
-  !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
-  int> = 0>
+template < class _Tp,
+   class _Up,
+   class _BinaryPredicate,
+   __enable_if_t<__desugars_to<_BinaryPredicate, 
std::equal_to<>>::value && !is_volatile<_Tp>::value &&
+ !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
+ int> = 0>
 _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
 __equal_iter_impl(_Tp* __first1, _Tp* __last1, _Up* __first2, 
_BinaryPredicate&) {
   return std::__constexpr_memcmp_equal(__first1, __first2, 
__element_count(__last1 - __first1));
@@ -94,12 +94,12 @@ template ::value && __is_identity<_Proj1>::value &&
+  __enable_if_t<__desugars_to<_Pred, std::equal_to<>>::value && 
__is_identity<_Proj1>::value &&
 __is_identity<_Proj2>::value && 
!is_volatile<_Tp>::value && !is_volatile<_Up>::value &&
 __libcpp_is_trivially_equality_comparable<_Tp, 
_Up>::value,
 int> = 0>
-_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool __equal_impl(
-_Tp* __first1, _Tp* __last1, _Up* __first2, _Up*, _Pred&, _Proj1&, 
_Proj2&) {
+_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
+

[flang] [compiler-rt] [libc] [libcxx] [clang-tools-extra] [llvm] [clang] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/68642

>From f0d93cc6a5cd485c654ab38221691db038bacc7d Mon Sep 17 00:00:00 2001
From: AntonRydahl 
Date: Mon, 9 Oct 2023 15:13:22 -0700
Subject: [PATCH 1/7] Merged __is_trivial_equality_predicate and
 __is_trivial_plus_operation into __desugars_to

---
 libcxx/include/CMakeLists.txt |  1 -
 libcxx/include/__algorithm/comp.h |  7 ++---
 libcxx/include/__algorithm/equal.h| 22 
 .../cpu_backends/transform_reduce.h   | 23 
 libcxx/include/__functional/operations.h  | 15 +--
 .../include/__functional/ranges_operations.h  |  7 ++---
 .../include/__numeric/pstl_transform_reduce.h |  2 +-
 .../include/__type_traits/operation_traits.h  |  4 +--
 .../include/__type_traits/predicate_traits.h  | 26 ---
 libcxx/include/module.modulemap.in|  1 -
 10 files changed, 40 insertions(+), 68 deletions(-)
 delete mode 100644 libcxx/include/__type_traits/predicate_traits.h

diff --git a/libcxx/include/CMakeLists.txt b/libcxx/include/CMakeLists.txt
index 340353f8ebb41c4..7eb09a06ccd482e 100644
--- a/libcxx/include/CMakeLists.txt
+++ b/libcxx/include/CMakeLists.txt
@@ -813,7 +813,6 @@ set(files
   __type_traits/negation.h
   __type_traits/noexcept_move_assign_container.h
   __type_traits/operation_traits.h
-  __type_traits/predicate_traits.h
   __type_traits/promote.h
   __type_traits/rank.h
   __type_traits/remove_all_extents.h
diff --git a/libcxx/include/__algorithm/comp.h 
b/libcxx/include/__algorithm/comp.h
index 9474536615ffb67..0993c37cce36a6b 100644
--- a/libcxx/include/__algorithm/comp.h
+++ b/libcxx/include/__algorithm/comp.h
@@ -10,8 +10,9 @@
 #define _LIBCPP___ALGORITHM_COMP_H
 
 #include <__config>
+#include <__functional/operations.h>
 #include <__type_traits/integral_constant.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
 #  pragma GCC system_header
@@ -26,8 +27,8 @@ struct __equal_to {
   }
 };
 
-template 
-struct __is_trivial_equality_predicate<__equal_to, _Lhs, _Rhs> : true_type {};
+template <>
+struct __desugars_to<__equal_to, std::equal_to<>> : true_type {};
 
 // The definition is required because __less is part of the ABI, but it's empty
 // because all comparisons should be transparent.
diff --git a/libcxx/include/__algorithm/equal.h 
b/libcxx/include/__algorithm/equal.h
index b69aeff92bb9289..35e82da15e4d058 100644
--- a/libcxx/include/__algorithm/equal.h
+++ b/libcxx/include/__algorithm/equal.h
@@ -15,6 +15,7 @@
 #include <__config>
 #include <__functional/identity.h>
 #include <__functional/invoke.h>
+#include <__functional/operations.h>
 #include <__iterator/distance.h>
 #include <__iterator/iterator_traits.h>
 #include <__string/constexpr_c_functions.h>
@@ -23,7 +24,7 @@
 #include <__type_traits/is_constant_evaluated.h>
 #include <__type_traits/is_equality_comparable.h>
 #include <__type_traits/is_volatile.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 #include <__utility/move.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
@@ -41,13 +42,12 @@ _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI 
_LIBCPP_CONSTEXPR_SINCE_CXX20 boo
   return true;
 }
 
-template <
-class _Tp,
-class _Up,
-class _BinaryPredicate,
-__enable_if_t<__is_trivial_equality_predicate<_BinaryPredicate, _Tp, 
_Up>::value && !is_volatile<_Tp>::value &&
-  !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
-  int> = 0>
+template < class _Tp,
+   class _Up,
+   class _BinaryPredicate,
+   __enable_if_t<__desugars_to<_BinaryPredicate, 
std::equal_to<>>::value && !is_volatile<_Tp>::value &&
+ !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
+ int> = 0>
 _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
 __equal_iter_impl(_Tp* __first1, _Tp* __last1, _Up* __first2, 
_BinaryPredicate&) {
   return std::__constexpr_memcmp_equal(__first1, __first2, 
__element_count(__last1 - __first1));
@@ -94,12 +94,12 @@ template ::value && __is_identity<_Proj1>::value &&
+  __enable_if_t<__desugars_to<_Pred, std::equal_to<>>::value && 
__is_identity<_Proj1>::value &&
 __is_identity<_Proj2>::value && 
!is_volatile<_Tp>::value && !is_volatile<_Up>::value &&
 __libcpp_is_trivially_equality_comparable<_Tp, 
_Up>::value,
 int> = 0>
-_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool __equal_impl(
-_Tp* __first1, _Tp* __last1, _Up* __first2, _Up*, _Pred&, _Proj1&, 
_Proj2&) {
+_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
+

[llvm] [mlir] [compiler-rt] [lldb] [clang] [libcxxabi] [libcxx] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

@@ -330,6 +330,23 @@ def getStdFlag(cfg, std):
 default=f"{shlex.quote(sys.executable)} 
{shlex.quote(str(Path(__file__).resolve().parent.parent.parent / 'run.py'))}",
 help="Custom executor to use instead of the configured default.",
 actions=lambda executor: [AddSubstitution("%{executor}", executor)],
+),
+Parameter(
+name="openmp_pstl_backend",
+choices=[True, False],
+type=bool,
+default=False,
+help="Enable the OpenMP compilation toolchain if the PSTL backend was 
set to OpenMP.",
+actions=lambda enabled: [
+AddCompileFlag("-fopenmp"),

AntonRydahl wrote:

I think there is unfortunately no better way of doing this at the moment, 
except for changing Clang. The problem is that OpenMP might already be shipped 
with the system, and then a lot of conflicts can happen if the wrong version of 
the OpenMP target library or the device runtime is picked up. But I agree that 
this is indeed not very user friendly and that it would be a lot nicer if Clang 
automatically did this.

https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang-tools-extra] [libc] [compiler-rt] [clang] [flang] [libcxx] [llvm] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/68642

>From f0d93cc6a5cd485c654ab38221691db038bacc7d Mon Sep 17 00:00:00 2001
From: AntonRydahl 
Date: Mon, 9 Oct 2023 15:13:22 -0700
Subject: [PATCH 1/8] Merged __is_trivial_equality_predicate and
 __is_trivial_plus_operation into __desugars_to

---
 libcxx/include/CMakeLists.txt |  1 -
 libcxx/include/__algorithm/comp.h |  7 ++---
 libcxx/include/__algorithm/equal.h| 22 
 .../cpu_backends/transform_reduce.h   | 23 
 libcxx/include/__functional/operations.h  | 15 +--
 .../include/__functional/ranges_operations.h  |  7 ++---
 .../include/__numeric/pstl_transform_reduce.h |  2 +-
 .../include/__type_traits/operation_traits.h  |  4 +--
 .../include/__type_traits/predicate_traits.h  | 26 ---
 libcxx/include/module.modulemap.in|  1 -
 10 files changed, 40 insertions(+), 68 deletions(-)
 delete mode 100644 libcxx/include/__type_traits/predicate_traits.h

diff --git a/libcxx/include/CMakeLists.txt b/libcxx/include/CMakeLists.txt
index 340353f8ebb41c4..7eb09a06ccd482e 100644
--- a/libcxx/include/CMakeLists.txt
+++ b/libcxx/include/CMakeLists.txt
@@ -813,7 +813,6 @@ set(files
   __type_traits/negation.h
   __type_traits/noexcept_move_assign_container.h
   __type_traits/operation_traits.h
-  __type_traits/predicate_traits.h
   __type_traits/promote.h
   __type_traits/rank.h
   __type_traits/remove_all_extents.h
diff --git a/libcxx/include/__algorithm/comp.h 
b/libcxx/include/__algorithm/comp.h
index 9474536615ffb67..0993c37cce36a6b 100644
--- a/libcxx/include/__algorithm/comp.h
+++ b/libcxx/include/__algorithm/comp.h
@@ -10,8 +10,9 @@
 #define _LIBCPP___ALGORITHM_COMP_H
 
 #include <__config>
+#include <__functional/operations.h>
 #include <__type_traits/integral_constant.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
 #  pragma GCC system_header
@@ -26,8 +27,8 @@ struct __equal_to {
   }
 };
 
-template 
-struct __is_trivial_equality_predicate<__equal_to, _Lhs, _Rhs> : true_type {};
+template <>
+struct __desugars_to<__equal_to, std::equal_to<>> : true_type {};
 
 // The definition is required because __less is part of the ABI, but it's empty
 // because all comparisons should be transparent.
diff --git a/libcxx/include/__algorithm/equal.h 
b/libcxx/include/__algorithm/equal.h
index b69aeff92bb9289..35e82da15e4d058 100644
--- a/libcxx/include/__algorithm/equal.h
+++ b/libcxx/include/__algorithm/equal.h
@@ -15,6 +15,7 @@
 #include <__config>
 #include <__functional/identity.h>
 #include <__functional/invoke.h>
+#include <__functional/operations.h>
 #include <__iterator/distance.h>
 #include <__iterator/iterator_traits.h>
 #include <__string/constexpr_c_functions.h>
@@ -23,7 +24,7 @@
 #include <__type_traits/is_constant_evaluated.h>
 #include <__type_traits/is_equality_comparable.h>
 #include <__type_traits/is_volatile.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 #include <__utility/move.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
@@ -41,13 +42,12 @@ _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI 
_LIBCPP_CONSTEXPR_SINCE_CXX20 boo
   return true;
 }
 
-template <
-class _Tp,
-class _Up,
-class _BinaryPredicate,
-__enable_if_t<__is_trivial_equality_predicate<_BinaryPredicate, _Tp, 
_Up>::value && !is_volatile<_Tp>::value &&
-  !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
-  int> = 0>
+template < class _Tp,
+   class _Up,
+   class _BinaryPredicate,
+   __enable_if_t<__desugars_to<_BinaryPredicate, 
std::equal_to<>>::value && !is_volatile<_Tp>::value &&
+ !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
+ int> = 0>
 _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
 __equal_iter_impl(_Tp* __first1, _Tp* __last1, _Up* __first2, 
_BinaryPredicate&) {
   return std::__constexpr_memcmp_equal(__first1, __first2, 
__element_count(__last1 - __first1));
@@ -94,12 +94,12 @@ template ::value && __is_identity<_Proj1>::value &&
+  __enable_if_t<__desugars_to<_Pred, std::equal_to<>>::value && 
__is_identity<_Proj1>::value &&
 __is_identity<_Proj2>::value && 
!is_volatile<_Tp>::value && !is_volatile<_Up>::value &&
 __libcpp_is_trivially_equality_comparable<_Tp, 
_Up>::value,
 int> = 0>
-_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool __equal_impl(
-_Tp* __first1, _Tp* __last1, _Up* __first2, _Up*, _Pred&, _Proj1&, 
_Proj2&) {
+_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
+

[libcxx] [clang] [compiler-rt] [llvm] [lldb] [libcxxabi] [mlir] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

@@ -466,6 +466,114 @@ Unpoisoning may not be an option, if (for example) you 
are not maintaining the a
 * You are using allocator, which does not call destructor during deallocation.
 * You are aware that memory allocated with an allocator may be accessed, even 
when unused by container.
 
+Offloading C++ Parallel Algorithms to GPUs
+--
+
+Experimental support for GPU offloading has been added to ``libc++``. The
+implementation uses OpenMP target offloading to leverage GPU compute resources.
+The OpenMP PSTL backend can target both NVIDIA and AMD GPUs.
+However, the implementation only supports contiguous iterators, such as 
+iterators for ``std::vector`` or ``std::array``.
+To enable the OpenMP offloading backend it must be selected with
+``LIBCXX_PSTL_BACKEND=openmp`` when installing ``libc++``. Further, when
+compiling a program, the user must specify the command line options
+``-fopenmp -fexperimental-library -stdlib=libc++``. To install LLVM with OpenMP
+offloading enabled, please read
+`the LLVM OpenMP FAQ. `_ 
+You may also want to to visit
+`the OpenMP offloading command-line argument reference. 
`_
 
+
+Example
+~~~
+
+The following is an example of offloading vector addition to a GPU using our
+standard library extension.
+
+.. code-block:: cpp
+
+  #include 
+  #include 
+
+  template
+  void axpy(const T1 a,std::vector& x, std::vector& y)
+  {
+std::transform(std::execution::par_unseq,x.begin(),x.end(), y.begin(), 
y.begin(),
+  [=](T2 xi, T3 yi){ return a*xi + yi; });
+  }
+
+The execution policy ``std::execution::par_unseq`` states that the algorithm's
+execution may be parallelized, vectorized, and migrated across threads. This is
+the only execution mode that is safe to offload to GPUs, and for all other
+execution modes the algorithms will execute on the CPU.
+Special attention must be paid to the lambda captures when enabling GPU
+offloading. If the lambda captures by reference, the user must manually map the
+variables to the device. If capturing by reference, the above example could
+be implemented in the following way.
+
+.. code-block:: cpp
+
+  template
+  void axpy(const T1 a,std::vector& x, std::vector& y)
+  {
+  # pragma omp target data map(to:a)
+std::transform(std::execution::par_unseq,x.begin(),x.end(), y.begin(), 
y.begin(),
+  [&](T2 xi, T3 yi){ return a*xi + yi; });
+  }
+
+However, if unified shared memory, USM, is enabled, no additional data mapping
+is necessary when capturing y reference.
+
+Compiling functions for GPUs with OpenMP
+
+
+The C++ standard defines that all accesses to memory are inside a single 
address
+space. However, discrete GPU systems have distinct address spaces. A single
+address space can be emulated if your system supports unified shared memory.
+However, many discrete GPU systems do not, and in those cases it is important 
to
+pass device function pointers to the parallel algorithms. Below is an example 
of
+how the OpenMP `declare target` directive can be used to mark that a function
+should be compiled for both host and device. The device address of a function
+pointer can be obtained with `target map(from:)`.
+
+.. code-block:: cpp
+
+  // Declare that the function must be compiled for both host and device
+  #pragma omp declare target
+  void cube(int& n) {n*=n*n; };
+  #pragma omp end declare target
+
+  int main()
+  {
+int * a =  new int[LEN];
+// Initialize the array to 2 on the device
+std::fill(std::execution::par_unseq,a, a+LEN,2);
+// Get the device pointer for cube
+void (*dcube)(int& n);
+#pragma omp target map(from:dcube)
+dcube = &cube;

AntonRydahl wrote:

@jdoerfert just made it possible to make the mapping automatically. If 
https://github.com/llvm/llvm-project/pull/71462 lands, we just need to use 
`#pragma omp declare target indirect to()`.

https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[mlir] [compiler-rt] [libcxx] [libcxxabi] [clang] [llvm] [lldb] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl edited 
https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang-tools-extra] [compiler-rt] [libcxx] [libc] [flang] [llvm] [clang] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/68642

>From f0d93cc6a5cd485c654ab38221691db038bacc7d Mon Sep 17 00:00:00 2001
From: AntonRydahl 
Date: Mon, 9 Oct 2023 15:13:22 -0700
Subject: [PATCH 1/9] Merged __is_trivial_equality_predicate and
 __is_trivial_plus_operation into __desugars_to

---
 libcxx/include/CMakeLists.txt |  1 -
 libcxx/include/__algorithm/comp.h |  7 ++---
 libcxx/include/__algorithm/equal.h| 22 
 .../cpu_backends/transform_reduce.h   | 23 
 libcxx/include/__functional/operations.h  | 15 +--
 .../include/__functional/ranges_operations.h  |  7 ++---
 .../include/__numeric/pstl_transform_reduce.h |  2 +-
 .../include/__type_traits/operation_traits.h  |  4 +--
 .../include/__type_traits/predicate_traits.h  | 26 ---
 libcxx/include/module.modulemap.in|  1 -
 10 files changed, 40 insertions(+), 68 deletions(-)
 delete mode 100644 libcxx/include/__type_traits/predicate_traits.h

diff --git a/libcxx/include/CMakeLists.txt b/libcxx/include/CMakeLists.txt
index 340353f8ebb41c4..7eb09a06ccd482e 100644
--- a/libcxx/include/CMakeLists.txt
+++ b/libcxx/include/CMakeLists.txt
@@ -813,7 +813,6 @@ set(files
   __type_traits/negation.h
   __type_traits/noexcept_move_assign_container.h
   __type_traits/operation_traits.h
-  __type_traits/predicate_traits.h
   __type_traits/promote.h
   __type_traits/rank.h
   __type_traits/remove_all_extents.h
diff --git a/libcxx/include/__algorithm/comp.h 
b/libcxx/include/__algorithm/comp.h
index 9474536615ffb67..0993c37cce36a6b 100644
--- a/libcxx/include/__algorithm/comp.h
+++ b/libcxx/include/__algorithm/comp.h
@@ -10,8 +10,9 @@
 #define _LIBCPP___ALGORITHM_COMP_H
 
 #include <__config>
+#include <__functional/operations.h>
 #include <__type_traits/integral_constant.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
 #  pragma GCC system_header
@@ -26,8 +27,8 @@ struct __equal_to {
   }
 };
 
-template 
-struct __is_trivial_equality_predicate<__equal_to, _Lhs, _Rhs> : true_type {};
+template <>
+struct __desugars_to<__equal_to, std::equal_to<>> : true_type {};
 
 // The definition is required because __less is part of the ABI, but it's empty
 // because all comparisons should be transparent.
diff --git a/libcxx/include/__algorithm/equal.h 
b/libcxx/include/__algorithm/equal.h
index b69aeff92bb9289..35e82da15e4d058 100644
--- a/libcxx/include/__algorithm/equal.h
+++ b/libcxx/include/__algorithm/equal.h
@@ -15,6 +15,7 @@
 #include <__config>
 #include <__functional/identity.h>
 #include <__functional/invoke.h>
+#include <__functional/operations.h>
 #include <__iterator/distance.h>
 #include <__iterator/iterator_traits.h>
 #include <__string/constexpr_c_functions.h>
@@ -23,7 +24,7 @@
 #include <__type_traits/is_constant_evaluated.h>
 #include <__type_traits/is_equality_comparable.h>
 #include <__type_traits/is_volatile.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 #include <__utility/move.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
@@ -41,13 +42,12 @@ _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI 
_LIBCPP_CONSTEXPR_SINCE_CXX20 boo
   return true;
 }
 
-template <
-class _Tp,
-class _Up,
-class _BinaryPredicate,
-__enable_if_t<__is_trivial_equality_predicate<_BinaryPredicate, _Tp, 
_Up>::value && !is_volatile<_Tp>::value &&
-  !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
-  int> = 0>
+template < class _Tp,
+   class _Up,
+   class _BinaryPredicate,
+   __enable_if_t<__desugars_to<_BinaryPredicate, 
std::equal_to<>>::value && !is_volatile<_Tp>::value &&
+ !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
+ int> = 0>
 _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
 __equal_iter_impl(_Tp* __first1, _Tp* __last1, _Up* __first2, 
_BinaryPredicate&) {
   return std::__constexpr_memcmp_equal(__first1, __first2, 
__element_count(__last1 - __first1));
@@ -94,12 +94,12 @@ template ::value && __is_identity<_Proj1>::value &&
+  __enable_if_t<__desugars_to<_Pred, std::equal_to<>>::value && 
__is_identity<_Proj1>::value &&
 __is_identity<_Proj2>::value && 
!is_volatile<_Tp>::value && !is_volatile<_Up>::value &&
 __libcpp_is_trivially_equality_comparable<_Tp, 
_Up>::value,
 int> = 0>
-_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool __equal_impl(
-_Tp* __first1, _Tp* __last1, _Up* __first2, _Up*, _Pred&, _Proj1&, 
_Proj2&) {
+_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
+

[libcxx] [libc] [clang] [clang-tools-extra] [llvm] [flang] [compiler-rt] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/68642

>From f0d93cc6a5cd485c654ab38221691db038bacc7d Mon Sep 17 00:00:00 2001
From: AntonRydahl 
Date: Mon, 9 Oct 2023 15:13:22 -0700
Subject: [PATCH 01/10] Merged __is_trivial_equality_predicate and
 __is_trivial_plus_operation into __desugars_to

---
 libcxx/include/CMakeLists.txt |  1 -
 libcxx/include/__algorithm/comp.h |  7 ++---
 libcxx/include/__algorithm/equal.h| 22 
 .../cpu_backends/transform_reduce.h   | 23 
 libcxx/include/__functional/operations.h  | 15 +--
 .../include/__functional/ranges_operations.h  |  7 ++---
 .../include/__numeric/pstl_transform_reduce.h |  2 +-
 .../include/__type_traits/operation_traits.h  |  4 +--
 .../include/__type_traits/predicate_traits.h  | 26 ---
 libcxx/include/module.modulemap.in|  1 -
 10 files changed, 40 insertions(+), 68 deletions(-)
 delete mode 100644 libcxx/include/__type_traits/predicate_traits.h

diff --git a/libcxx/include/CMakeLists.txt b/libcxx/include/CMakeLists.txt
index 340353f8ebb41c4..7eb09a06ccd482e 100644
--- a/libcxx/include/CMakeLists.txt
+++ b/libcxx/include/CMakeLists.txt
@@ -813,7 +813,6 @@ set(files
   __type_traits/negation.h
   __type_traits/noexcept_move_assign_container.h
   __type_traits/operation_traits.h
-  __type_traits/predicate_traits.h
   __type_traits/promote.h
   __type_traits/rank.h
   __type_traits/remove_all_extents.h
diff --git a/libcxx/include/__algorithm/comp.h 
b/libcxx/include/__algorithm/comp.h
index 9474536615ffb67..0993c37cce36a6b 100644
--- a/libcxx/include/__algorithm/comp.h
+++ b/libcxx/include/__algorithm/comp.h
@@ -10,8 +10,9 @@
 #define _LIBCPP___ALGORITHM_COMP_H
 
 #include <__config>
+#include <__functional/operations.h>
 #include <__type_traits/integral_constant.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
 #  pragma GCC system_header
@@ -26,8 +27,8 @@ struct __equal_to {
   }
 };
 
-template 
-struct __is_trivial_equality_predicate<__equal_to, _Lhs, _Rhs> : true_type {};
+template <>
+struct __desugars_to<__equal_to, std::equal_to<>> : true_type {};
 
 // The definition is required because __less is part of the ABI, but it's empty
 // because all comparisons should be transparent.
diff --git a/libcxx/include/__algorithm/equal.h 
b/libcxx/include/__algorithm/equal.h
index b69aeff92bb9289..35e82da15e4d058 100644
--- a/libcxx/include/__algorithm/equal.h
+++ b/libcxx/include/__algorithm/equal.h
@@ -15,6 +15,7 @@
 #include <__config>
 #include <__functional/identity.h>
 #include <__functional/invoke.h>
+#include <__functional/operations.h>
 #include <__iterator/distance.h>
 #include <__iterator/iterator_traits.h>
 #include <__string/constexpr_c_functions.h>
@@ -23,7 +24,7 @@
 #include <__type_traits/is_constant_evaluated.h>
 #include <__type_traits/is_equality_comparable.h>
 #include <__type_traits/is_volatile.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 #include <__utility/move.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
@@ -41,13 +42,12 @@ _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI 
_LIBCPP_CONSTEXPR_SINCE_CXX20 boo
   return true;
 }
 
-template <
-class _Tp,
-class _Up,
-class _BinaryPredicate,
-__enable_if_t<__is_trivial_equality_predicate<_BinaryPredicate, _Tp, 
_Up>::value && !is_volatile<_Tp>::value &&
-  !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
-  int> = 0>
+template < class _Tp,
+   class _Up,
+   class _BinaryPredicate,
+   __enable_if_t<__desugars_to<_BinaryPredicate, 
std::equal_to<>>::value && !is_volatile<_Tp>::value &&
+ !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
+ int> = 0>
 _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
 __equal_iter_impl(_Tp* __first1, _Tp* __last1, _Up* __first2, 
_BinaryPredicate&) {
   return std::__constexpr_memcmp_equal(__first1, __first2, 
__element_count(__last1 - __first1));
@@ -94,12 +94,12 @@ template ::value && __is_identity<_Proj1>::value &&
+  __enable_if_t<__desugars_to<_Pred, std::equal_to<>>::value && 
__is_identity<_Proj1>::value &&
 __is_identity<_Proj2>::value && 
!is_volatile<_Tp>::value && !is_volatile<_Up>::value &&
 __libcpp_is_trivially_equality_comparable<_Tp, 
_Up>::value,
 int> = 0>
-_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool __equal_impl(
-_Tp* __first1, _Tp* __last1, _Up* __first2, _Up*, _Pred&, _Proj1&, 
_Proj2&) {
+_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool

[llvm] [compiler-rt] [clang] [flang] [libc] [libcxx] [clang-tools-extra] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/68642

>From f0d93cc6a5cd485c654ab38221691db038bacc7d Mon Sep 17 00:00:00 2001
From: AntonRydahl 
Date: Mon, 9 Oct 2023 15:13:22 -0700
Subject: [PATCH 01/11] Merged __is_trivial_equality_predicate and
 __is_trivial_plus_operation into __desugars_to

---
 libcxx/include/CMakeLists.txt |  1 -
 libcxx/include/__algorithm/comp.h |  7 ++---
 libcxx/include/__algorithm/equal.h| 22 
 .../cpu_backends/transform_reduce.h   | 23 
 libcxx/include/__functional/operations.h  | 15 +--
 .../include/__functional/ranges_operations.h  |  7 ++---
 .../include/__numeric/pstl_transform_reduce.h |  2 +-
 .../include/__type_traits/operation_traits.h  |  4 +--
 .../include/__type_traits/predicate_traits.h  | 26 ---
 libcxx/include/module.modulemap.in|  1 -
 10 files changed, 40 insertions(+), 68 deletions(-)
 delete mode 100644 libcxx/include/__type_traits/predicate_traits.h

diff --git a/libcxx/include/CMakeLists.txt b/libcxx/include/CMakeLists.txt
index 340353f8ebb41c4..7eb09a06ccd482e 100644
--- a/libcxx/include/CMakeLists.txt
+++ b/libcxx/include/CMakeLists.txt
@@ -813,7 +813,6 @@ set(files
   __type_traits/negation.h
   __type_traits/noexcept_move_assign_container.h
   __type_traits/operation_traits.h
-  __type_traits/predicate_traits.h
   __type_traits/promote.h
   __type_traits/rank.h
   __type_traits/remove_all_extents.h
diff --git a/libcxx/include/__algorithm/comp.h 
b/libcxx/include/__algorithm/comp.h
index 9474536615ffb67..0993c37cce36a6b 100644
--- a/libcxx/include/__algorithm/comp.h
+++ b/libcxx/include/__algorithm/comp.h
@@ -10,8 +10,9 @@
 #define _LIBCPP___ALGORITHM_COMP_H
 
 #include <__config>
+#include <__functional/operations.h>
 #include <__type_traits/integral_constant.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
 #  pragma GCC system_header
@@ -26,8 +27,8 @@ struct __equal_to {
   }
 };
 
-template 
-struct __is_trivial_equality_predicate<__equal_to, _Lhs, _Rhs> : true_type {};
+template <>
+struct __desugars_to<__equal_to, std::equal_to<>> : true_type {};
 
 // The definition is required because __less is part of the ABI, but it's empty
 // because all comparisons should be transparent.
diff --git a/libcxx/include/__algorithm/equal.h 
b/libcxx/include/__algorithm/equal.h
index b69aeff92bb9289..35e82da15e4d058 100644
--- a/libcxx/include/__algorithm/equal.h
+++ b/libcxx/include/__algorithm/equal.h
@@ -15,6 +15,7 @@
 #include <__config>
 #include <__functional/identity.h>
 #include <__functional/invoke.h>
+#include <__functional/operations.h>
 #include <__iterator/distance.h>
 #include <__iterator/iterator_traits.h>
 #include <__string/constexpr_c_functions.h>
@@ -23,7 +24,7 @@
 #include <__type_traits/is_constant_evaluated.h>
 #include <__type_traits/is_equality_comparable.h>
 #include <__type_traits/is_volatile.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 #include <__utility/move.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
@@ -41,13 +42,12 @@ _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI 
_LIBCPP_CONSTEXPR_SINCE_CXX20 boo
   return true;
 }
 
-template <
-class _Tp,
-class _Up,
-class _BinaryPredicate,
-__enable_if_t<__is_trivial_equality_predicate<_BinaryPredicate, _Tp, 
_Up>::value && !is_volatile<_Tp>::value &&
-  !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
-  int> = 0>
+template < class _Tp,
+   class _Up,
+   class _BinaryPredicate,
+   __enable_if_t<__desugars_to<_BinaryPredicate, 
std::equal_to<>>::value && !is_volatile<_Tp>::value &&
+ !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
+ int> = 0>
 _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
 __equal_iter_impl(_Tp* __first1, _Tp* __last1, _Up* __first2, 
_BinaryPredicate&) {
   return std::__constexpr_memcmp_equal(__first1, __first2, 
__element_count(__last1 - __first1));
@@ -94,12 +94,12 @@ template ::value && __is_identity<_Proj1>::value &&
+  __enable_if_t<__desugars_to<_Pred, std::equal_to<>>::value && 
__is_identity<_Proj1>::value &&
 __is_identity<_Proj2>::value && 
!is_volatile<_Tp>::value && !is_volatile<_Up>::value &&
 __libcpp_is_trivially_equality_comparable<_Tp, 
_Up>::value,
 int> = 0>
-_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool __equal_impl(
-_Tp* __first1, _Tp* __last1, _Up* __first2, _Up*, _Pred&, _Proj1&, 
_Proj2&) {
+_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool

[compiler-rt] [flang] [clang-tools-extra] [llvm] [libcxx] [clang] [libc] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/68642

>From f0d93cc6a5cd485c654ab38221691db038bacc7d Mon Sep 17 00:00:00 2001
From: AntonRydahl 
Date: Mon, 9 Oct 2023 15:13:22 -0700
Subject: [PATCH 01/12] Merged __is_trivial_equality_predicate and
 __is_trivial_plus_operation into __desugars_to

---
 libcxx/include/CMakeLists.txt |  1 -
 libcxx/include/__algorithm/comp.h |  7 ++---
 libcxx/include/__algorithm/equal.h| 22 
 .../cpu_backends/transform_reduce.h   | 23 
 libcxx/include/__functional/operations.h  | 15 +--
 .../include/__functional/ranges_operations.h  |  7 ++---
 .../include/__numeric/pstl_transform_reduce.h |  2 +-
 .../include/__type_traits/operation_traits.h  |  4 +--
 .../include/__type_traits/predicate_traits.h  | 26 ---
 libcxx/include/module.modulemap.in|  1 -
 10 files changed, 40 insertions(+), 68 deletions(-)
 delete mode 100644 libcxx/include/__type_traits/predicate_traits.h

diff --git a/libcxx/include/CMakeLists.txt b/libcxx/include/CMakeLists.txt
index 340353f8ebb41c4..7eb09a06ccd482e 100644
--- a/libcxx/include/CMakeLists.txt
+++ b/libcxx/include/CMakeLists.txt
@@ -813,7 +813,6 @@ set(files
   __type_traits/negation.h
   __type_traits/noexcept_move_assign_container.h
   __type_traits/operation_traits.h
-  __type_traits/predicate_traits.h
   __type_traits/promote.h
   __type_traits/rank.h
   __type_traits/remove_all_extents.h
diff --git a/libcxx/include/__algorithm/comp.h 
b/libcxx/include/__algorithm/comp.h
index 9474536615ffb67..0993c37cce36a6b 100644
--- a/libcxx/include/__algorithm/comp.h
+++ b/libcxx/include/__algorithm/comp.h
@@ -10,8 +10,9 @@
 #define _LIBCPP___ALGORITHM_COMP_H
 
 #include <__config>
+#include <__functional/operations.h>
 #include <__type_traits/integral_constant.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
 #  pragma GCC system_header
@@ -26,8 +27,8 @@ struct __equal_to {
   }
 };
 
-template 
-struct __is_trivial_equality_predicate<__equal_to, _Lhs, _Rhs> : true_type {};
+template <>
+struct __desugars_to<__equal_to, std::equal_to<>> : true_type {};
 
 // The definition is required because __less is part of the ABI, but it's empty
 // because all comparisons should be transparent.
diff --git a/libcxx/include/__algorithm/equal.h 
b/libcxx/include/__algorithm/equal.h
index b69aeff92bb9289..35e82da15e4d058 100644
--- a/libcxx/include/__algorithm/equal.h
+++ b/libcxx/include/__algorithm/equal.h
@@ -15,6 +15,7 @@
 #include <__config>
 #include <__functional/identity.h>
 #include <__functional/invoke.h>
+#include <__functional/operations.h>
 #include <__iterator/distance.h>
 #include <__iterator/iterator_traits.h>
 #include <__string/constexpr_c_functions.h>
@@ -23,7 +24,7 @@
 #include <__type_traits/is_constant_evaluated.h>
 #include <__type_traits/is_equality_comparable.h>
 #include <__type_traits/is_volatile.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 #include <__utility/move.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
@@ -41,13 +42,12 @@ _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI 
_LIBCPP_CONSTEXPR_SINCE_CXX20 boo
   return true;
 }
 
-template <
-class _Tp,
-class _Up,
-class _BinaryPredicate,
-__enable_if_t<__is_trivial_equality_predicate<_BinaryPredicate, _Tp, 
_Up>::value && !is_volatile<_Tp>::value &&
-  !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
-  int> = 0>
+template < class _Tp,
+   class _Up,
+   class _BinaryPredicate,
+   __enable_if_t<__desugars_to<_BinaryPredicate, 
std::equal_to<>>::value && !is_volatile<_Tp>::value &&
+ !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
+ int> = 0>
 _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
 __equal_iter_impl(_Tp* __first1, _Tp* __last1, _Up* __first2, 
_BinaryPredicate&) {
   return std::__constexpr_memcmp_equal(__first1, __first2, 
__element_count(__last1 - __first1));
@@ -94,12 +94,12 @@ template ::value && __is_identity<_Proj1>::value &&
+  __enable_if_t<__desugars_to<_Pred, std::equal_to<>>::value && 
__is_identity<_Proj1>::value &&
 __is_identity<_Proj2>::value && 
!is_volatile<_Tp>::value && !is_volatile<_Up>::value &&
 __libcpp_is_trivially_equality_comparable<_Tp, 
_Up>::value,
 int> = 0>
-_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool __equal_impl(
-_Tp* __first1, _Tp* __last1, _Up* __first2, _Up*, _Pred&, _Proj1&, 
_Proj2&) {
+_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool

[libc] [compiler-rt] [clang-tools-extra] [flang] [libcxx] [llvm] [clang] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/68642

>From f0d93cc6a5cd485c654ab38221691db038bacc7d Mon Sep 17 00:00:00 2001
From: AntonRydahl 
Date: Mon, 9 Oct 2023 15:13:22 -0700
Subject: [PATCH 01/13] Merged __is_trivial_equality_predicate and
 __is_trivial_plus_operation into __desugars_to

---
 libcxx/include/CMakeLists.txt |  1 -
 libcxx/include/__algorithm/comp.h |  7 ++---
 libcxx/include/__algorithm/equal.h| 22 
 .../cpu_backends/transform_reduce.h   | 23 
 libcxx/include/__functional/operations.h  | 15 +--
 .../include/__functional/ranges_operations.h  |  7 ++---
 .../include/__numeric/pstl_transform_reduce.h |  2 +-
 .../include/__type_traits/operation_traits.h  |  4 +--
 .../include/__type_traits/predicate_traits.h  | 26 ---
 libcxx/include/module.modulemap.in|  1 -
 10 files changed, 40 insertions(+), 68 deletions(-)
 delete mode 100644 libcxx/include/__type_traits/predicate_traits.h

diff --git a/libcxx/include/CMakeLists.txt b/libcxx/include/CMakeLists.txt
index 340353f8ebb41c4..7eb09a06ccd482e 100644
--- a/libcxx/include/CMakeLists.txt
+++ b/libcxx/include/CMakeLists.txt
@@ -813,7 +813,6 @@ set(files
   __type_traits/negation.h
   __type_traits/noexcept_move_assign_container.h
   __type_traits/operation_traits.h
-  __type_traits/predicate_traits.h
   __type_traits/promote.h
   __type_traits/rank.h
   __type_traits/remove_all_extents.h
diff --git a/libcxx/include/__algorithm/comp.h 
b/libcxx/include/__algorithm/comp.h
index 9474536615ffb67..0993c37cce36a6b 100644
--- a/libcxx/include/__algorithm/comp.h
+++ b/libcxx/include/__algorithm/comp.h
@@ -10,8 +10,9 @@
 #define _LIBCPP___ALGORITHM_COMP_H
 
 #include <__config>
+#include <__functional/operations.h>
 #include <__type_traits/integral_constant.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
 #  pragma GCC system_header
@@ -26,8 +27,8 @@ struct __equal_to {
   }
 };
 
-template 
-struct __is_trivial_equality_predicate<__equal_to, _Lhs, _Rhs> : true_type {};
+template <>
+struct __desugars_to<__equal_to, std::equal_to<>> : true_type {};
 
 // The definition is required because __less is part of the ABI, but it's empty
 // because all comparisons should be transparent.
diff --git a/libcxx/include/__algorithm/equal.h 
b/libcxx/include/__algorithm/equal.h
index b69aeff92bb9289..35e82da15e4d058 100644
--- a/libcxx/include/__algorithm/equal.h
+++ b/libcxx/include/__algorithm/equal.h
@@ -15,6 +15,7 @@
 #include <__config>
 #include <__functional/identity.h>
 #include <__functional/invoke.h>
+#include <__functional/operations.h>
 #include <__iterator/distance.h>
 #include <__iterator/iterator_traits.h>
 #include <__string/constexpr_c_functions.h>
@@ -23,7 +24,7 @@
 #include <__type_traits/is_constant_evaluated.h>
 #include <__type_traits/is_equality_comparable.h>
 #include <__type_traits/is_volatile.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 #include <__utility/move.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
@@ -41,13 +42,12 @@ _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI 
_LIBCPP_CONSTEXPR_SINCE_CXX20 boo
   return true;
 }
 
-template <
-class _Tp,
-class _Up,
-class _BinaryPredicate,
-__enable_if_t<__is_trivial_equality_predicate<_BinaryPredicate, _Tp, 
_Up>::value && !is_volatile<_Tp>::value &&
-  !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
-  int> = 0>
+template < class _Tp,
+   class _Up,
+   class _BinaryPredicate,
+   __enable_if_t<__desugars_to<_BinaryPredicate, 
std::equal_to<>>::value && !is_volatile<_Tp>::value &&
+ !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
+ int> = 0>
 _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
 __equal_iter_impl(_Tp* __first1, _Tp* __last1, _Up* __first2, 
_BinaryPredicate&) {
   return std::__constexpr_memcmp_equal(__first1, __first2, 
__element_count(__last1 - __first1));
@@ -94,12 +94,12 @@ template ::value && __is_identity<_Proj1>::value &&
+  __enable_if_t<__desugars_to<_Pred, std::equal_to<>>::value && 
__is_identity<_Proj1>::value &&
 __is_identity<_Proj2>::value && 
!is_volatile<_Tp>::value && !is_volatile<_Up>::value &&
 __libcpp_is_trivially_equality_comparable<_Tp, 
_Up>::value,
 int> = 0>
-_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool __equal_impl(
-_Tp* __first1, _Tp* __last1, _Up* __first2, _Up*, _Pred&, _Proj1&, 
_Proj2&) {
+_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool

[clang-tools-extra] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

@@ -330,6 +330,23 @@ def getStdFlag(cfg, std):
 default=f"{shlex.quote(sys.executable)} 
{shlex.quote(str(Path(__file__).resolve().parent.parent.parent / 'run.py'))}",
 help="Custom executor to use instead of the configured default.",
 actions=lambda executor: [AddSubstitution("%{executor}", executor)],
+),
+Parameter(
+name="openmp_pstl_backend",
+choices=[True, False],
+type=bool,
+default=False,
+help="Enable the OpenMP compilation toolchain if the PSTL backend was 
set to OpenMP.",
+actions=lambda enabled: [
+AddCompileFlag("-fopenmp"),
+# The linker needs to find the correct version of libomptarget
+AddLinkFlag("-Wl,-rpath,%{lib}/../../lib"),
+AddLinkFlag("-L%{lib}/../../lib"),
+#  The preprocessor needs to find the omp.h header
+AddFlag("-I 
%{lib}/../../runtimes/runtimes-bins/openmp/runtime/src"),
+# If the OpenMP PSTL backend was enbaled, we wish to run the tests 
for it

AntonRydahl wrote:

Hi @jhuber6 and @jdoerfert. @ldionne and I just discussed how to modify the 
`libc++` test configuration to be able to offload to GPUs. I could not get it 
to work without appending linking and include paths. Do you think `clang` 
should automatically look for `libomptarget.so` and `omp.h` in the correct 
places, or is it really necessary to add the paths?

https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[mlir] [lldb] [libcxxabi] [clang] [compiler-rt] [llvm] [libcxx] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

@@ -466,6 +466,114 @@ Unpoisoning may not be an option, if (for example) you 
are not maintaining the a
 * You are using allocator, which does not call destructor during deallocation.
 * You are aware that memory allocated with an allocator may be accessed, even 
when unused by container.
 
+Offloading C++ Parallel Algorithms to GPUs
+--
+
+Experimental support for GPU offloading has been added to ``libc++``. The
+implementation uses OpenMP target offloading to leverage GPU compute resources.
+The OpenMP PSTL backend can target both NVIDIA and AMD GPUs.
+However, the implementation only supports contiguous iterators, such as 
+iterators for ``std::vector`` or ``std::array``.
+To enable the OpenMP offloading backend it must be selected with
+``LIBCXX_PSTL_BACKEND=openmp`` when installing ``libc++``. Further, when
+compiling a program, the user must specify the command line options
+``-fopenmp -fexperimental-library -stdlib=libc++``. To install LLVM with OpenMP
+offloading enabled, please read
+`the LLVM OpenMP FAQ. `_ 
+You may also want to to visit
+`the OpenMP offloading command-line argument reference. 
`_
 
+
+Example
+~~~
+
+The following is an example of offloading vector addition to a GPU using our
+standard library extension.
+
+.. code-block:: cpp
+
+  #include 
+  #include 
+
+  template
+  void axpy(const T1 a,std::vector& x, std::vector& y)
+  {
+std::transform(std::execution::par_unseq,x.begin(),x.end(), y.begin(), 
y.begin(),
+  [=](T2 xi, T3 yi){ return a*xi + yi; });
+  }
+
+The execution policy ``std::execution::par_unseq`` states that the algorithm's
+execution may be parallelized, vectorized, and migrated across threads. This is
+the only execution mode that is safe to offload to GPUs, and for all other
+execution modes the algorithms will execute on the CPU.
+Special attention must be paid to the lambda captures when enabling GPU
+offloading. If the lambda captures by reference, the user must manually map the
+variables to the device. If capturing by reference, the above example could
+be implemented in the following way.
+
+.. code-block:: cpp
+
+  template
+  void axpy(const T1 a,std::vector& x, std::vector& y)
+  {
+  # pragma omp target data map(to:a)
+std::transform(std::execution::par_unseq,x.begin(),x.end(), y.begin(), 
y.begin(),

AntonRydahl wrote:

I think they are correct, but I will const qualify the x vector. It is reading 
from x and y and writing to y, just like the classical BLAS AXPY routine. That 
corresponds to 4) from https://en.cppreference.com/w/cpp/algorithm/transform.

https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang-tools-extra] [libc] [libcxx] [compiler-rt] [flang] [clang] [llvm] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

@@ -18,8 +18,11 @@
 
 _LIBCPP_BEGIN_NAMESPACE_STD
 
-template 
-struct __is_trivial_plus_operation : false_type {};
+template 

AntonRydahl wrote:

Hi @ldionne!

I am unsure about what I am supposed to do with the following:
```C++
template  struct __desugars_to<__equal_tag, __equal, _Tp, 
_Up> : true_type {};
```
Do we want to match the function from `include/__algorithm/equal.h`?

https://github.com/llvm/llvm-project/pull/68642
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[libcxx] [clang] [libc] [flang] [llvm] [compiler-rt] [clang-tools-extra] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/68642

>From f0d93cc6a5cd485c654ab38221691db038bacc7d Mon Sep 17 00:00:00 2001
From: AntonRydahl 
Date: Mon, 9 Oct 2023 15:13:22 -0700
Subject: [PATCH 1/5] Merged __is_trivial_equality_predicate and
 __is_trivial_plus_operation into __desugars_to

---
 libcxx/include/CMakeLists.txt |  1 -
 libcxx/include/__algorithm/comp.h |  7 ++---
 libcxx/include/__algorithm/equal.h| 22 
 .../cpu_backends/transform_reduce.h   | 23 
 libcxx/include/__functional/operations.h  | 15 +--
 .../include/__functional/ranges_operations.h  |  7 ++---
 .../include/__numeric/pstl_transform_reduce.h |  2 +-
 .../include/__type_traits/operation_traits.h  |  4 +--
 .../include/__type_traits/predicate_traits.h  | 26 ---
 libcxx/include/module.modulemap.in|  1 -
 10 files changed, 40 insertions(+), 68 deletions(-)
 delete mode 100644 libcxx/include/__type_traits/predicate_traits.h

diff --git a/libcxx/include/CMakeLists.txt b/libcxx/include/CMakeLists.txt
index 340353f8ebb41c4..7eb09a06ccd482e 100644
--- a/libcxx/include/CMakeLists.txt
+++ b/libcxx/include/CMakeLists.txt
@@ -813,7 +813,6 @@ set(files
   __type_traits/negation.h
   __type_traits/noexcept_move_assign_container.h
   __type_traits/operation_traits.h
-  __type_traits/predicate_traits.h
   __type_traits/promote.h
   __type_traits/rank.h
   __type_traits/remove_all_extents.h
diff --git a/libcxx/include/__algorithm/comp.h 
b/libcxx/include/__algorithm/comp.h
index 9474536615ffb67..0993c37cce36a6b 100644
--- a/libcxx/include/__algorithm/comp.h
+++ b/libcxx/include/__algorithm/comp.h
@@ -10,8 +10,9 @@
 #define _LIBCPP___ALGORITHM_COMP_H
 
 #include <__config>
+#include <__functional/operations.h>
 #include <__type_traits/integral_constant.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
 #  pragma GCC system_header
@@ -26,8 +27,8 @@ struct __equal_to {
   }
 };
 
-template 
-struct __is_trivial_equality_predicate<__equal_to, _Lhs, _Rhs> : true_type {};
+template <>
+struct __desugars_to<__equal_to, std::equal_to<>> : true_type {};
 
 // The definition is required because __less is part of the ABI, but it's empty
 // because all comparisons should be transparent.
diff --git a/libcxx/include/__algorithm/equal.h 
b/libcxx/include/__algorithm/equal.h
index b69aeff92bb9289..35e82da15e4d058 100644
--- a/libcxx/include/__algorithm/equal.h
+++ b/libcxx/include/__algorithm/equal.h
@@ -15,6 +15,7 @@
 #include <__config>
 #include <__functional/identity.h>
 #include <__functional/invoke.h>
+#include <__functional/operations.h>
 #include <__iterator/distance.h>
 #include <__iterator/iterator_traits.h>
 #include <__string/constexpr_c_functions.h>
@@ -23,7 +24,7 @@
 #include <__type_traits/is_constant_evaluated.h>
 #include <__type_traits/is_equality_comparable.h>
 #include <__type_traits/is_volatile.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 #include <__utility/move.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
@@ -41,13 +42,12 @@ _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI 
_LIBCPP_CONSTEXPR_SINCE_CXX20 boo
   return true;
 }
 
-template <
-class _Tp,
-class _Up,
-class _BinaryPredicate,
-__enable_if_t<__is_trivial_equality_predicate<_BinaryPredicate, _Tp, 
_Up>::value && !is_volatile<_Tp>::value &&
-  !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
-  int> = 0>
+template < class _Tp,
+   class _Up,
+   class _BinaryPredicate,
+   __enable_if_t<__desugars_to<_BinaryPredicate, 
std::equal_to<>>::value && !is_volatile<_Tp>::value &&
+ !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
+ int> = 0>
 _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
 __equal_iter_impl(_Tp* __first1, _Tp* __last1, _Up* __first2, 
_BinaryPredicate&) {
   return std::__constexpr_memcmp_equal(__first1, __first2, 
__element_count(__last1 - __first1));
@@ -94,12 +94,12 @@ template ::value && __is_identity<_Proj1>::value &&
+  __enable_if_t<__desugars_to<_Pred, std::equal_to<>>::value && 
__is_identity<_Proj1>::value &&
 __is_identity<_Proj2>::value && 
!is_volatile<_Tp>::value && !is_volatile<_Up>::value &&
 __libcpp_is_trivially_equality_comparable<_Tp, 
_Up>::value,
 int> = 0>
-_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool __equal_impl(
-_Tp* __first1, _Tp* __last1, _Up* __first2, _Up*, _Pred&, _Proj1&, 
_Proj2&) {
+_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
+

[compiler-rt] [libcxx] [clang] [libc] [flang] [llvm] [clang-tools-extra] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/68642

>From f0d93cc6a5cd485c654ab38221691db038bacc7d Mon Sep 17 00:00:00 2001
From: AntonRydahl 
Date: Mon, 9 Oct 2023 15:13:22 -0700
Subject: [PATCH 1/5] Merged __is_trivial_equality_predicate and
 __is_trivial_plus_operation into __desugars_to

---
 libcxx/include/CMakeLists.txt |  1 -
 libcxx/include/__algorithm/comp.h |  7 ++---
 libcxx/include/__algorithm/equal.h| 22 
 .../cpu_backends/transform_reduce.h   | 23 
 libcxx/include/__functional/operations.h  | 15 +--
 .../include/__functional/ranges_operations.h  |  7 ++---
 .../include/__numeric/pstl_transform_reduce.h |  2 +-
 .../include/__type_traits/operation_traits.h  |  4 +--
 .../include/__type_traits/predicate_traits.h  | 26 ---
 libcxx/include/module.modulemap.in|  1 -
 10 files changed, 40 insertions(+), 68 deletions(-)
 delete mode 100644 libcxx/include/__type_traits/predicate_traits.h

diff --git a/libcxx/include/CMakeLists.txt b/libcxx/include/CMakeLists.txt
index 340353f8ebb41c4..7eb09a06ccd482e 100644
--- a/libcxx/include/CMakeLists.txt
+++ b/libcxx/include/CMakeLists.txt
@@ -813,7 +813,6 @@ set(files
   __type_traits/negation.h
   __type_traits/noexcept_move_assign_container.h
   __type_traits/operation_traits.h
-  __type_traits/predicate_traits.h
   __type_traits/promote.h
   __type_traits/rank.h
   __type_traits/remove_all_extents.h
diff --git a/libcxx/include/__algorithm/comp.h 
b/libcxx/include/__algorithm/comp.h
index 9474536615ffb67..0993c37cce36a6b 100644
--- a/libcxx/include/__algorithm/comp.h
+++ b/libcxx/include/__algorithm/comp.h
@@ -10,8 +10,9 @@
 #define _LIBCPP___ALGORITHM_COMP_H
 
 #include <__config>
+#include <__functional/operations.h>
 #include <__type_traits/integral_constant.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
 #  pragma GCC system_header
@@ -26,8 +27,8 @@ struct __equal_to {
   }
 };
 
-template 
-struct __is_trivial_equality_predicate<__equal_to, _Lhs, _Rhs> : true_type {};
+template <>
+struct __desugars_to<__equal_to, std::equal_to<>> : true_type {};
 
 // The definition is required because __less is part of the ABI, but it's empty
 // because all comparisons should be transparent.
diff --git a/libcxx/include/__algorithm/equal.h 
b/libcxx/include/__algorithm/equal.h
index b69aeff92bb9289..35e82da15e4d058 100644
--- a/libcxx/include/__algorithm/equal.h
+++ b/libcxx/include/__algorithm/equal.h
@@ -15,6 +15,7 @@
 #include <__config>
 #include <__functional/identity.h>
 #include <__functional/invoke.h>
+#include <__functional/operations.h>
 #include <__iterator/distance.h>
 #include <__iterator/iterator_traits.h>
 #include <__string/constexpr_c_functions.h>
@@ -23,7 +24,7 @@
 #include <__type_traits/is_constant_evaluated.h>
 #include <__type_traits/is_equality_comparable.h>
 #include <__type_traits/is_volatile.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 #include <__utility/move.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
@@ -41,13 +42,12 @@ _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI 
_LIBCPP_CONSTEXPR_SINCE_CXX20 boo
   return true;
 }
 
-template <
-class _Tp,
-class _Up,
-class _BinaryPredicate,
-__enable_if_t<__is_trivial_equality_predicate<_BinaryPredicate, _Tp, 
_Up>::value && !is_volatile<_Tp>::value &&
-  !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
-  int> = 0>
+template < class _Tp,
+   class _Up,
+   class _BinaryPredicate,
+   __enable_if_t<__desugars_to<_BinaryPredicate, 
std::equal_to<>>::value && !is_volatile<_Tp>::value &&
+ !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
+ int> = 0>
 _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
 __equal_iter_impl(_Tp* __first1, _Tp* __last1, _Up* __first2, 
_BinaryPredicate&) {
   return std::__constexpr_memcmp_equal(__first1, __first2, 
__element_count(__last1 - __first1));
@@ -94,12 +94,12 @@ template ::value && __is_identity<_Proj1>::value &&
+  __enable_if_t<__desugars_to<_Pred, std::equal_to<>>::value && 
__is_identity<_Proj1>::value &&
 __is_identity<_Proj2>::value && 
!is_volatile<_Tp>::value && !is_volatile<_Up>::value &&
 __libcpp_is_trivially_equality_comparable<_Tp, 
_Up>::value,
 int> = 0>
-_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool __equal_impl(
-_Tp* __first1, _Tp* __last1, _Up* __first2, _Up*, _Pred&, _Proj1&, 
_Proj2&) {
+_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
+

[clang-tools-extra] [libc][libm][GPU] Add missing vendor entrypoints to the GPU version of `libm` (PR #66034)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/66034

>From 7a357059d3ca78afcb02e33a6afce225029a3836 Mon Sep 17 00:00:00 2001
From: antonrydahl 
Date: Mon, 11 Sep 2023 17:06:41 -0700
Subject: [PATCH 1/3] [libc][libm][GPU] Add missing vendor entrypoints to the
 GPU version of `libm`

There are a number of entry-points in `libm` for which no target-agnostic 
implementations exist and the built-ins do not lower correctly on either AMDGPU 
or NVPTX targets. This patch adds the vendor entry-points as a temporary 
solution.
---
 libc/config/gpu/entrypoints.txt   |  37 -
 libc/src/math/CMakeLists.txt  |  15 ++
 libc/src/math/acos.h  |  18 +++
 libc/src/math/acosh.h |  18 +++
 libc/src/math/asin.h  |  18 +++
 libc/src/math/asinh.h |  18 +++
 libc/src/math/atan.h  |  18 +++
 libc/src/math/atan2.h |  18 +++
 libc/src/math/atan2f.h|  18 +++
 libc/src/math/atanh.h |  18 +++
 libc/src/math/erf.h   |  18 +++
 libc/src/math/expm1.h |  18 +++
 libc/src/math/gpu/vendor/CMakeLists.txt   | 143 ++
 libc/src/math/gpu/vendor/acos.cpp |  18 +++
 libc/src/math/gpu/vendor/acosh.cpp|  18 +++
 libc/src/math/gpu/vendor/amdgpu/amdgpu.h  |  25 +++
 .../src/math/gpu/vendor/amdgpu/declarations.h |  34 -
 libc/src/math/gpu/vendor/asin.cpp |  18 +++
 libc/src/math/gpu/vendor/asinh.cpp|  18 +++
 libc/src/math/gpu/vendor/atan.cpp |  18 +++
 libc/src/math/gpu/vendor/atan2.cpp|  20 +++
 libc/src/math/gpu/vendor/atan2f.cpp   |  21 +++
 libc/src/math/gpu/vendor/atanh.cpp|  18 +++
 libc/src/math/gpu/vendor/erf.cpp  |  18 +++
 libc/src/math/gpu/vendor/erff.cpp |  18 +++
 libc/src/math/gpu/vendor/exp.cpp  |  18 +++
 libc/src/math/gpu/vendor/exp10.cpp|  18 +++
 libc/src/math/gpu/vendor/exp2.cpp |  18 +++
 libc/src/math/gpu/vendor/expm1.cpp|  18 +++
 libc/src/math/gpu/vendor/log.cpp  |  18 +++
 libc/src/math/gpu/vendor/log10.cpp|  18 +++
 libc/src/math/gpu/vendor/log10f.cpp   |  19 +++
 libc/src/math/gpu/vendor/log1p.cpp|  18 +++
 libc/src/math/gpu/vendor/log1pf.cpp   |  18 +++
 libc/src/math/gpu/vendor/log2.cpp |  18 +++
 libc/src/math/gpu/vendor/log2f.cpp|  18 +++
 libc/src/math/gpu/vendor/logb.cpp |  18 +++
 libc/src/math/gpu/vendor/logbf.cpp|  19 +++
 libc/src/math/gpu/vendor/logf.cpp |  18 +++
 libc/src/math/gpu/vendor/lrint.cpp|  18 +++
 libc/src/math/gpu/vendor/lrintf.cpp   |  18 +++
 libc/src/math/gpu/vendor/lround.cpp   |  18 +++
 libc/src/math/gpu/vendor/lroundf.cpp  |  20 +++
 libc/src/math/gpu/vendor/nvptx/declarations.h |  25 +++
 libc/src/math/gpu/vendor/nvptx/nvptx.h|  25 +++
 libc/src/math/gpu/vendor/tgamma.cpp   |  18 +++
 libc/src/math/gpu/vendor/tgammaf.cpp  |  18 +++
 libc/src/math/sincos.h|  18 +++
 libc/src/math/tgamma.h|  18 +++
 libc/src/math/tgammaf.h   |  18 +++
 50 files changed, 1081 insertions(+), 6 deletions(-)
 create mode 100644 libc/src/math/acos.h
 create mode 100644 libc/src/math/acosh.h
 create mode 100644 libc/src/math/asin.h
 create mode 100644 libc/src/math/asinh.h
 create mode 100644 libc/src/math/atan.h
 create mode 100644 libc/src/math/atan2.h
 create mode 100644 libc/src/math/atan2f.h
 create mode 100644 libc/src/math/atanh.h
 create mode 100644 libc/src/math/erf.h
 create mode 100644 libc/src/math/expm1.h
 create mode 100644 libc/src/math/gpu/vendor/acos.cpp
 create mode 100644 libc/src/math/gpu/vendor/acosh.cpp
 create mode 100644 libc/src/math/gpu/vendor/asin.cpp
 create mode 100644 libc/src/math/gpu/vendor/asinh.cpp
 create mode 100644 libc/src/math/gpu/vendor/atan.cpp
 create mode 100644 libc/src/math/gpu/vendor/atan2.cpp
 create mode 100644 libc/src/math/gpu/vendor/atan2f.cpp
 create mode 100644 libc/src/math/gpu/vendor/atanh.cpp
 create mode 100644 libc/src/math/gpu/vendor/erf.cpp
 create mode 100644 libc/src/math/gpu/vendor/erff.cpp
 create mode 100644 libc/src/math/gpu/vendor/exp.cpp
 create mode 100644 libc/src/math/gpu/vendor/exp10.cpp
 create mode 100644 libc/src/math/gpu/vendor/exp2.cpp
 create mode 100644 libc/src/math/gpu/vendor/expm1.cpp
 create mode 100644 libc/src/math/gpu/vendor/log.cpp
 create mode 100644 libc/src/math/gpu/vendor/log10.cpp
 create mode 100644 libc/src/math/gpu/vendor/log10f.cpp
 create mode 100644 libc/src/math/gpu/vendor/log1p.cpp
 create mode 100644 libc/src/math/gpu/vendor/log1pf.cpp
 create mode 100644 libc/src/math/gpu/vendor/log2.cpp
 creat

[clang-tools-extra] [libc][libm][GPU] Add missing vendor entrypoints to the GPU version of `libm` (PR #66034)

[Anton Rydahl via cfe-commits]

AntonRydahl wrote:

> LG after fixing the header formatting.

Thanks a bunch! I don't know how I did not notice the misformatted headers. I 
have fixed those now.

https://github.com/llvm/llvm-project/pull/66034
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [libc][libm][GPU] Add missing vendor entrypoints to the GPU version of `libm` (PR #66034)

[Anton Rydahl via cfe-commits]

AntonRydahl wrote:

> LG after fixing the header formatting.

Thanks a bunch! I don't know how I did not notice the misformatted headers. I 
have fixed those now.

https://github.com/llvm/llvm-project/pull/66034
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [libc][libm][GPU] Add missing vendor entrypoints to the GPU version of `libm` (PR #66034)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl closed 
https://github.com/llvm/llvm-project/pull/66034
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang-tools-extra] [libc][libm][GPU] Add missing vendor entrypoints to the GPU version of `libm` (PR #66034)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl closed 
https://github.com/llvm/llvm-project/pull/66034
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

@@ -0,0 +1,142 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM 
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--===//
+
+#ifndef _LIBCPP___ALGORITHM_PSTL_BACKENDS_OPENMP_BACKEND_TRANSFORM_H
+#define _LIBCPP___ALGORITHM_PSTL_BACKENDS_OPENMP_BACKEND_TRANSFORM_H
+
+#include <__algorithm/pstl_backends/cpu_backends/backend.h>
+#include <__algorithm/pstl_backends/openmp/backend.h>
+#include <__algorithm/transform.h>
+#include <__config>
+#include <__iterator/concepts.h>
+#include <__type_traits/is_execution_policy.h>
+#include 
+
+#if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
+#  pragma GCC system_header
+#endif
+
+#if !defined(_LIBCPP_HAS_NO_INCOMPLETE_PSTL) && _LIBCPP_STD_VER >= 17
+
+_LIBCPP_BEGIN_NAMESPACE_STD
+
+namespace __par_backend {
+inline namespace __omp_gpu_backend {
+
+//===--===//
+// Templates for two iterators
+//===--===//
+
+template 
+_LIBCPP_HIDE_FROM_ABI optional<__empty>
+__omp_parallel_for_simd(_Tp* __first1, _DifferenceType __n, _Up* __first2, 
_Function __f) noexcept {

AntonRydahl wrote:

`__in`, `__out`, and `__inout` are reserved keywords in `libc++` so I have 
renamed the variables `__in1`, `__out1`, and so on instead. I hope that is ok. 

https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

@@ -0,0 +1,142 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM 
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--===//
+
+#ifndef _LIBCPP___ALGORITHM_PSTL_BACKENDS_OPENMP_BACKEND_TRANSFORM_H
+#define _LIBCPP___ALGORITHM_PSTL_BACKENDS_OPENMP_BACKEND_TRANSFORM_H
+
+#include <__algorithm/pstl_backends/cpu_backends/backend.h>
+#include <__algorithm/pstl_backends/openmp/backend.h>
+#include <__algorithm/transform.h>
+#include <__config>
+#include <__iterator/concepts.h>
+#include <__type_traits/is_execution_policy.h>
+#include 
+
+#if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
+#  pragma GCC system_header
+#endif
+
+#if !defined(_LIBCPP_HAS_NO_INCOMPLETE_PSTL) && _LIBCPP_STD_VER >= 17
+
+_LIBCPP_BEGIN_NAMESPACE_STD
+
+namespace __par_backend {
+inline namespace __omp_gpu_backend {
+
+//===--===//
+// Templates for two iterators
+//===--===//
+
+template 
+_LIBCPP_HIDE_FROM_ABI optional<__empty>
+__omp_parallel_for_simd(_Tp* __first1, _DifferenceType __n, _Up* __first2, 
_Function __f) noexcept {
+  __omp_gpu_backend::__omp_map_alloc(__first2, __n);
+  __omp_gpu_backend::__omp_map_to(__first1, __n);
+#  pragma omp target teams distribute parallel for simd
+  for (_DifferenceType __i = 0; __i < __n; ++__i)
+*(__first2 + __i) = __f(*(__first1 + __i));
+  __omp_gpu_backend::__omp_map_from(__first2, __n);
+  __omp_gpu_backend::__omp_map_free(__first1, __n);
+  return __empty{};
+}
+
+// Extracting the underlying pointer
+
+template 
+_LIBCPP_HIDE_FROM_ABI optional<__empty>
+__parallel_for_simd(_Iterator1 __first1, _DifferenceType __n, _Iterator2 
__first2, _Function __f) noexcept {
+  return __omp_gpu_backend::__omp_parallel_for_simd(
+  __omp_gpu_backend::__omp_extract_base_ptr(__first1),
+  __n,
+  __omp_gpu_backend::__omp_extract_base_ptr(__first2),
+  __f);
+}
+
+//===--===//
+// Templates for three iterator
+//===--===//
+
+template 
+_LIBCPP_HIDE_FROM_ABI optional<__empty>
+__omp_parallel_for_simd(_Tp* __first1, _DifferenceType __n, _Up* __first2, 
_Vp* __first3, _Function __f) noexcept {

AntonRydahl wrote:

I have renamed all the functions `__omp_`. 

https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang-tools-extra] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

@@ -0,0 +1,73 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM 
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--===//
+
+#ifndef _LIBCPP___ALGORITHM_PSTL_BACKENDS_OPENMP_BACKEND_FILL_H

AntonRydahl wrote:

After some additional consideration, I think we should not reuse `for_each`, 
because we can eliminate the host to device data transfer when making a 
specialized implementation of `fill`.

https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

@@ -0,0 +1,73 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM 
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--===//
+
+#ifndef _LIBCPP___ALGORITHM_PSTL_BACKENDS_OPENMP_BACKEND_FILL_H

AntonRydahl wrote:

After some additional consideration, I think we should not reuse `for_each`, 
because we can eliminate the host to device data transfer when making a 
specialized implementation of `fill`.

https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang-tools-extra] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

@@ -0,0 +1,349 @@
+//===--===//

AntonRydahl wrote:

I have written a section about this in the `libc++` documentation. Do you think 
the description is sufficiently detailed?

https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

@@ -0,0 +1,349 @@
+//===--===//

AntonRydahl wrote:

I have written a section about this in the `libc++` documentation. Do you think 
the description is sufficiently detailed?

https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

AntonRydahl wrote:

@ldionne do you think it looks right at this point? 😄 

https://github.com/llvm/llvm-project/pull/68642
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang-tools-extra] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

AntonRydahl wrote:

@ldionne do you think it looks right at this point? 😄 

https://github.com/llvm/llvm-project/pull/68642
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl edited 
https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang-tools-extra] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

AntonRydahl wrote:

When adding the logic to error when if OpenMP is not enabled, I had to add a 
bit of CMake logic to allow installing the library without getting the error. 
On the good side, this also made it possible to enable the OpenMP toolchain 
when running the LIT tests, if the OpenMP backend was selected.

I now found that a few of the LIT tests failed, because I had not thought of 
handling cases where the base type of the iterator is not trivially copyable. 
For instance, some of the LIT tests may throw exceptions in the copy 
constructor. The commit I pushed today makes those tests pass while still 
allowing my local tests to be executed on the GPU.

What do you think of this change?

https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/66968

>From b35340e47de896c9933c54ce617538c46cf01488 Mon Sep 17 00:00:00 2001
From: AntonRydahl 
Date: Wed, 20 Sep 2023 17:06:10 -0700
Subject: [PATCH 1/5] Adding OpenMP Offloading Backend for C++ Parallel
 Algorithms

---
 libcxx/CMakeLists.txt | 14 +++
 libcxx/include/CMakeLists.txt |  5 +
 libcxx/include/__algorithm/pstl_backend.h |  8 ++
 .../__algorithm/pstl_backends/gpu_backend.h   | 21 +
 .../pstl_backends/gpu_backends/backend.h  | 33 +++
 .../pstl_backends/gpu_backends/fill.h | 59 
 .../pstl_backends/gpu_backends/for_each.h | 59 
 .../pstl_backends/gpu_backends/omp_offload.h  | 91 +++
 libcxx/include/__config_site.in   |  1 +
 9 files changed, 291 insertions(+)
 create mode 100644 libcxx/include/__algorithm/pstl_backends/gpu_backend.h
 create mode 100644 
libcxx/include/__algorithm/pstl_backends/gpu_backends/backend.h
 create mode 100644 libcxx/include/__algorithm/pstl_backends/gpu_backends/fill.h
 create mode 100644 
libcxx/include/__algorithm/pstl_backends/gpu_backends/for_each.h
 create mode 100644 
libcxx/include/__algorithm/pstl_backends/gpu_backends/omp_offload.h

diff --git a/libcxx/CMakeLists.txt b/libcxx/CMakeLists.txt
index bb2898b799bcef9..43d2a448de79584 100644
--- a/libcxx/CMakeLists.txt
+++ b/libcxx/CMakeLists.txt
@@ -290,6 +290,8 @@ option(LIBCXX_HAS_WIN32_THREAD_API "Ignore auto-detection 
and force use of win32
 option(LIBCXX_HAS_EXTERNAL_THREAD_API
   "Build libc++ with an externalized threading API.
This option may only be set to ON when LIBCXX_ENABLE_THREADS=ON." OFF)
+option(LIBCXX_ENABLE_GPU_OFFLOAD 
+  "Build libc++ with support for GPU offload" OFF)
 
 if (LIBCXX_ENABLE_THREADS)
   set(LIBCXX_PSTL_CPU_BACKEND "std_thread" CACHE STRING "Which PSTL CPU 
backend to use")
@@ -297,6 +299,14 @@ else()
   set(LIBCXX_PSTL_CPU_BACKEND "serial" CACHE STRING "Which PSTL CPU backend to 
use")
 endif()
 
+if (NOT DEFINED LIBCXX_PSTL_GPU_BACKEND)
+  if (${LIBCXX_ENABLE_GPU_OFFLOAD})
+set(LIBCXX_PSTL_GPU_BACKEND "omp_offload" CACHE STRING "Which PSTL GPU 
backend to use")
+  else()
+set(LIBCXX_PSTL_GPU_BACKEND "none" CACHE STRING "Which PSTL GPU backend to 
use")
+  endif()
+endif()
+
 # Misc options 
 # FIXME: Turn -pedantic back ON. It is currently off because it warns
 # about #include_next which is used everywhere.
@@ -809,6 +819,10 @@ else()
Valid backends are: serial, std_thread and libdispatch")
 endif()
 
+if (LIBCXX_PSTL_GPU_BACKEND STREQUAL "omp_offload")
+  config_define(1 _LIBCPP_PSTL_GPU_BACKEND_OMP_OFFLOAD)
+endif()
+
 if (LIBCXX_ABI_DEFINES)
   set(abi_defines)
   foreach (abi_define ${LIBCXX_ABI_DEFINES})
diff --git a/libcxx/include/CMakeLists.txt b/libcxx/include/CMakeLists.txt
index 2ec755236dbaee2..a3d72df61a86dde 100644
--- a/libcxx/include/CMakeLists.txt
+++ b/libcxx/include/CMakeLists.txt
@@ -85,6 +85,11 @@ set(files
   __algorithm/pstl_backends/cpu_backends/thread.h
   __algorithm/pstl_backends/cpu_backends/transform.h
   __algorithm/pstl_backends/cpu_backends/transform_reduce.h
+  __algorithm/pstl_backends/gpu_backend.h
+  __algorithm/pstl_backends/gpu_backends/backend.h
+  __algorithm/pstl_backends/gpu_backends/fill.h
+  __algorithm/pstl_backends/gpu_backends/for_each.h
+  __algorithm/pstl_backends/gpu_backends/omp_offload.h
   __algorithm/pstl_copy.h
   __algorithm/pstl_count.h
   __algorithm/pstl_fill.h
diff --git a/libcxx/include/__algorithm/pstl_backend.h 
b/libcxx/include/__algorithm/pstl_backend.h
index 93372f019031b63..f051e0ce9be13c3 100644
--- a/libcxx/include/__algorithm/pstl_backend.h
+++ b/libcxx/include/__algorithm/pstl_backend.h
@@ -10,6 +10,7 @@
 #define _LIBCPP___ALGORITHM_PSTL_BACKEND_H
 
 #include <__algorithm/pstl_backends/cpu_backend.h>
+#include <__algorithm/pstl_backends/gpu_backend.h>
 #include <__config>
 #include 
 
@@ -179,10 +180,17 @@ struct __select_backend {
   using type = __cpu_backend_tag;
 };
 
+#   if defined(_LIBCPP_PSTL_GPU_BACKEND_OMP_OFFLOAD)
+template <>
+struct __select_backend {
+  using type = __gpu_backend_tag;
+};
+#   else
 template <>
 struct __select_backend {
   using type = __cpu_backend_tag;
 };
+#   endif
 
 #  else
 
diff --git a/libcxx/include/__algorithm/pstl_backends/gpu_backend.h 
b/libcxx/include/__algorithm/pstl_backends/gpu_backend.h
new file mode 100644
index 000..46a85f77b5deb99
--- /dev/null
+++ b/libcxx/include/__algorithm/pstl_backends/gpu_backend.h
@@ -0,0 +1,21 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM 
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--

[clang-tools-extra] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/66968

>From b35340e47de896c9933c54ce617538c46cf01488 Mon Sep 17 00:00:00 2001
From: AntonRydahl 
Date: Wed, 20 Sep 2023 17:06:10 -0700
Subject: [PATCH 1/5] Adding OpenMP Offloading Backend for C++ Parallel
 Algorithms

---
 libcxx/CMakeLists.txt | 14 +++
 libcxx/include/CMakeLists.txt |  5 +
 libcxx/include/__algorithm/pstl_backend.h |  8 ++
 .../__algorithm/pstl_backends/gpu_backend.h   | 21 +
 .../pstl_backends/gpu_backends/backend.h  | 33 +++
 .../pstl_backends/gpu_backends/fill.h | 59 
 .../pstl_backends/gpu_backends/for_each.h | 59 
 .../pstl_backends/gpu_backends/omp_offload.h  | 91 +++
 libcxx/include/__config_site.in   |  1 +
 9 files changed, 291 insertions(+)
 create mode 100644 libcxx/include/__algorithm/pstl_backends/gpu_backend.h
 create mode 100644 
libcxx/include/__algorithm/pstl_backends/gpu_backends/backend.h
 create mode 100644 libcxx/include/__algorithm/pstl_backends/gpu_backends/fill.h
 create mode 100644 
libcxx/include/__algorithm/pstl_backends/gpu_backends/for_each.h
 create mode 100644 
libcxx/include/__algorithm/pstl_backends/gpu_backends/omp_offload.h

diff --git a/libcxx/CMakeLists.txt b/libcxx/CMakeLists.txt
index bb2898b799bcef9..43d2a448de79584 100644
--- a/libcxx/CMakeLists.txt
+++ b/libcxx/CMakeLists.txt
@@ -290,6 +290,8 @@ option(LIBCXX_HAS_WIN32_THREAD_API "Ignore auto-detection 
and force use of win32
 option(LIBCXX_HAS_EXTERNAL_THREAD_API
   "Build libc++ with an externalized threading API.
This option may only be set to ON when LIBCXX_ENABLE_THREADS=ON." OFF)
+option(LIBCXX_ENABLE_GPU_OFFLOAD 
+  "Build libc++ with support for GPU offload" OFF)
 
 if (LIBCXX_ENABLE_THREADS)
   set(LIBCXX_PSTL_CPU_BACKEND "std_thread" CACHE STRING "Which PSTL CPU 
backend to use")
@@ -297,6 +299,14 @@ else()
   set(LIBCXX_PSTL_CPU_BACKEND "serial" CACHE STRING "Which PSTL CPU backend to 
use")
 endif()
 
+if (NOT DEFINED LIBCXX_PSTL_GPU_BACKEND)
+  if (${LIBCXX_ENABLE_GPU_OFFLOAD})
+set(LIBCXX_PSTL_GPU_BACKEND "omp_offload" CACHE STRING "Which PSTL GPU 
backend to use")
+  else()
+set(LIBCXX_PSTL_GPU_BACKEND "none" CACHE STRING "Which PSTL GPU backend to 
use")
+  endif()
+endif()
+
 # Misc options 
 # FIXME: Turn -pedantic back ON. It is currently off because it warns
 # about #include_next which is used everywhere.
@@ -809,6 +819,10 @@ else()
Valid backends are: serial, std_thread and libdispatch")
 endif()
 
+if (LIBCXX_PSTL_GPU_BACKEND STREQUAL "omp_offload")
+  config_define(1 _LIBCPP_PSTL_GPU_BACKEND_OMP_OFFLOAD)
+endif()
+
 if (LIBCXX_ABI_DEFINES)
   set(abi_defines)
   foreach (abi_define ${LIBCXX_ABI_DEFINES})
diff --git a/libcxx/include/CMakeLists.txt b/libcxx/include/CMakeLists.txt
index 2ec755236dbaee2..a3d72df61a86dde 100644
--- a/libcxx/include/CMakeLists.txt
+++ b/libcxx/include/CMakeLists.txt
@@ -85,6 +85,11 @@ set(files
   __algorithm/pstl_backends/cpu_backends/thread.h
   __algorithm/pstl_backends/cpu_backends/transform.h
   __algorithm/pstl_backends/cpu_backends/transform_reduce.h
+  __algorithm/pstl_backends/gpu_backend.h
+  __algorithm/pstl_backends/gpu_backends/backend.h
+  __algorithm/pstl_backends/gpu_backends/fill.h
+  __algorithm/pstl_backends/gpu_backends/for_each.h
+  __algorithm/pstl_backends/gpu_backends/omp_offload.h
   __algorithm/pstl_copy.h
   __algorithm/pstl_count.h
   __algorithm/pstl_fill.h
diff --git a/libcxx/include/__algorithm/pstl_backend.h 
b/libcxx/include/__algorithm/pstl_backend.h
index 93372f019031b63..f051e0ce9be13c3 100644
--- a/libcxx/include/__algorithm/pstl_backend.h
+++ b/libcxx/include/__algorithm/pstl_backend.h
@@ -10,6 +10,7 @@
 #define _LIBCPP___ALGORITHM_PSTL_BACKEND_H
 
 #include <__algorithm/pstl_backends/cpu_backend.h>
+#include <__algorithm/pstl_backends/gpu_backend.h>
 #include <__config>
 #include 
 
@@ -179,10 +180,17 @@ struct __select_backend {
   using type = __cpu_backend_tag;
 };
 
+#   if defined(_LIBCPP_PSTL_GPU_BACKEND_OMP_OFFLOAD)
+template <>
+struct __select_backend {
+  using type = __gpu_backend_tag;
+};
+#   else
 template <>
 struct __select_backend {
   using type = __cpu_backend_tag;
 };
+#   endif
 
 #  else
 
diff --git a/libcxx/include/__algorithm/pstl_backends/gpu_backend.h 
b/libcxx/include/__algorithm/pstl_backends/gpu_backend.h
new file mode 100644
index 000..46a85f77b5deb99
--- /dev/null
+++ b/libcxx/include/__algorithm/pstl_backends/gpu_backend.h
@@ -0,0 +1,21 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM 
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--

[clang] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/66968

>From b35340e47de896c9933c54ce617538c46cf01488 Mon Sep 17 00:00:00 2001
From: AntonRydahl 
Date: Wed, 20 Sep 2023 17:06:10 -0700
Subject: [PATCH 1/6] Adding OpenMP Offloading Backend for C++ Parallel
 Algorithms

---
 libcxx/CMakeLists.txt | 14 +++
 libcxx/include/CMakeLists.txt |  5 +
 libcxx/include/__algorithm/pstl_backend.h |  8 ++
 .../__algorithm/pstl_backends/gpu_backend.h   | 21 +
 .../pstl_backends/gpu_backends/backend.h  | 33 +++
 .../pstl_backends/gpu_backends/fill.h | 59 
 .../pstl_backends/gpu_backends/for_each.h | 59 
 .../pstl_backends/gpu_backends/omp_offload.h  | 91 +++
 libcxx/include/__config_site.in   |  1 +
 9 files changed, 291 insertions(+)
 create mode 100644 libcxx/include/__algorithm/pstl_backends/gpu_backend.h
 create mode 100644 
libcxx/include/__algorithm/pstl_backends/gpu_backends/backend.h
 create mode 100644 libcxx/include/__algorithm/pstl_backends/gpu_backends/fill.h
 create mode 100644 
libcxx/include/__algorithm/pstl_backends/gpu_backends/for_each.h
 create mode 100644 
libcxx/include/__algorithm/pstl_backends/gpu_backends/omp_offload.h

diff --git a/libcxx/CMakeLists.txt b/libcxx/CMakeLists.txt
index bb2898b799bcef9..43d2a448de79584 100644
--- a/libcxx/CMakeLists.txt
+++ b/libcxx/CMakeLists.txt
@@ -290,6 +290,8 @@ option(LIBCXX_HAS_WIN32_THREAD_API "Ignore auto-detection 
and force use of win32
 option(LIBCXX_HAS_EXTERNAL_THREAD_API
   "Build libc++ with an externalized threading API.
This option may only be set to ON when LIBCXX_ENABLE_THREADS=ON." OFF)
+option(LIBCXX_ENABLE_GPU_OFFLOAD 
+  "Build libc++ with support for GPU offload" OFF)
 
 if (LIBCXX_ENABLE_THREADS)
   set(LIBCXX_PSTL_CPU_BACKEND "std_thread" CACHE STRING "Which PSTL CPU 
backend to use")
@@ -297,6 +299,14 @@ else()
   set(LIBCXX_PSTL_CPU_BACKEND "serial" CACHE STRING "Which PSTL CPU backend to 
use")
 endif()
 
+if (NOT DEFINED LIBCXX_PSTL_GPU_BACKEND)
+  if (${LIBCXX_ENABLE_GPU_OFFLOAD})
+set(LIBCXX_PSTL_GPU_BACKEND "omp_offload" CACHE STRING "Which PSTL GPU 
backend to use")
+  else()
+set(LIBCXX_PSTL_GPU_BACKEND "none" CACHE STRING "Which PSTL GPU backend to 
use")
+  endif()
+endif()
+
 # Misc options 
 # FIXME: Turn -pedantic back ON. It is currently off because it warns
 # about #include_next which is used everywhere.
@@ -809,6 +819,10 @@ else()
Valid backends are: serial, std_thread and libdispatch")
 endif()
 
+if (LIBCXX_PSTL_GPU_BACKEND STREQUAL "omp_offload")
+  config_define(1 _LIBCPP_PSTL_GPU_BACKEND_OMP_OFFLOAD)
+endif()
+
 if (LIBCXX_ABI_DEFINES)
   set(abi_defines)
   foreach (abi_define ${LIBCXX_ABI_DEFINES})
diff --git a/libcxx/include/CMakeLists.txt b/libcxx/include/CMakeLists.txt
index 2ec755236dbaee2..a3d72df61a86dde 100644
--- a/libcxx/include/CMakeLists.txt
+++ b/libcxx/include/CMakeLists.txt
@@ -85,6 +85,11 @@ set(files
   __algorithm/pstl_backends/cpu_backends/thread.h
   __algorithm/pstl_backends/cpu_backends/transform.h
   __algorithm/pstl_backends/cpu_backends/transform_reduce.h
+  __algorithm/pstl_backends/gpu_backend.h
+  __algorithm/pstl_backends/gpu_backends/backend.h
+  __algorithm/pstl_backends/gpu_backends/fill.h
+  __algorithm/pstl_backends/gpu_backends/for_each.h
+  __algorithm/pstl_backends/gpu_backends/omp_offload.h
   __algorithm/pstl_copy.h
   __algorithm/pstl_count.h
   __algorithm/pstl_fill.h
diff --git a/libcxx/include/__algorithm/pstl_backend.h 
b/libcxx/include/__algorithm/pstl_backend.h
index 93372f019031b63..f051e0ce9be13c3 100644
--- a/libcxx/include/__algorithm/pstl_backend.h
+++ b/libcxx/include/__algorithm/pstl_backend.h
@@ -10,6 +10,7 @@
 #define _LIBCPP___ALGORITHM_PSTL_BACKEND_H
 
 #include <__algorithm/pstl_backends/cpu_backend.h>
+#include <__algorithm/pstl_backends/gpu_backend.h>
 #include <__config>
 #include 
 
@@ -179,10 +180,17 @@ struct __select_backend {
   using type = __cpu_backend_tag;
 };
 
+#   if defined(_LIBCPP_PSTL_GPU_BACKEND_OMP_OFFLOAD)
+template <>
+struct __select_backend {
+  using type = __gpu_backend_tag;
+};
+#   else
 template <>
 struct __select_backend {
   using type = __cpu_backend_tag;
 };
+#   endif
 
 #  else
 
diff --git a/libcxx/include/__algorithm/pstl_backends/gpu_backend.h 
b/libcxx/include/__algorithm/pstl_backends/gpu_backend.h
new file mode 100644
index 000..46a85f77b5deb99
--- /dev/null
+++ b/libcxx/include/__algorithm/pstl_backends/gpu_backend.h
@@ -0,0 +1,21 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM 
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--

[clang-tools-extra] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/66968

>From b35340e47de896c9933c54ce617538c46cf01488 Mon Sep 17 00:00:00 2001
From: AntonRydahl 
Date: Wed, 20 Sep 2023 17:06:10 -0700
Subject: [PATCH 1/6] Adding OpenMP Offloading Backend for C++ Parallel
 Algorithms

---
 libcxx/CMakeLists.txt | 14 +++
 libcxx/include/CMakeLists.txt |  5 +
 libcxx/include/__algorithm/pstl_backend.h |  8 ++
 .../__algorithm/pstl_backends/gpu_backend.h   | 21 +
 .../pstl_backends/gpu_backends/backend.h  | 33 +++
 .../pstl_backends/gpu_backends/fill.h | 59 
 .../pstl_backends/gpu_backends/for_each.h | 59 
 .../pstl_backends/gpu_backends/omp_offload.h  | 91 +++
 libcxx/include/__config_site.in   |  1 +
 9 files changed, 291 insertions(+)
 create mode 100644 libcxx/include/__algorithm/pstl_backends/gpu_backend.h
 create mode 100644 
libcxx/include/__algorithm/pstl_backends/gpu_backends/backend.h
 create mode 100644 libcxx/include/__algorithm/pstl_backends/gpu_backends/fill.h
 create mode 100644 
libcxx/include/__algorithm/pstl_backends/gpu_backends/for_each.h
 create mode 100644 
libcxx/include/__algorithm/pstl_backends/gpu_backends/omp_offload.h

diff --git a/libcxx/CMakeLists.txt b/libcxx/CMakeLists.txt
index bb2898b799bcef9..43d2a448de79584 100644
--- a/libcxx/CMakeLists.txt
+++ b/libcxx/CMakeLists.txt
@@ -290,6 +290,8 @@ option(LIBCXX_HAS_WIN32_THREAD_API "Ignore auto-detection 
and force use of win32
 option(LIBCXX_HAS_EXTERNAL_THREAD_API
   "Build libc++ with an externalized threading API.
This option may only be set to ON when LIBCXX_ENABLE_THREADS=ON." OFF)
+option(LIBCXX_ENABLE_GPU_OFFLOAD 
+  "Build libc++ with support for GPU offload" OFF)
 
 if (LIBCXX_ENABLE_THREADS)
   set(LIBCXX_PSTL_CPU_BACKEND "std_thread" CACHE STRING "Which PSTL CPU 
backend to use")
@@ -297,6 +299,14 @@ else()
   set(LIBCXX_PSTL_CPU_BACKEND "serial" CACHE STRING "Which PSTL CPU backend to 
use")
 endif()
 
+if (NOT DEFINED LIBCXX_PSTL_GPU_BACKEND)
+  if (${LIBCXX_ENABLE_GPU_OFFLOAD})
+set(LIBCXX_PSTL_GPU_BACKEND "omp_offload" CACHE STRING "Which PSTL GPU 
backend to use")
+  else()
+set(LIBCXX_PSTL_GPU_BACKEND "none" CACHE STRING "Which PSTL GPU backend to 
use")
+  endif()
+endif()
+
 # Misc options 
 # FIXME: Turn -pedantic back ON. It is currently off because it warns
 # about #include_next which is used everywhere.
@@ -809,6 +819,10 @@ else()
Valid backends are: serial, std_thread and libdispatch")
 endif()
 
+if (LIBCXX_PSTL_GPU_BACKEND STREQUAL "omp_offload")
+  config_define(1 _LIBCPP_PSTL_GPU_BACKEND_OMP_OFFLOAD)
+endif()
+
 if (LIBCXX_ABI_DEFINES)
   set(abi_defines)
   foreach (abi_define ${LIBCXX_ABI_DEFINES})
diff --git a/libcxx/include/CMakeLists.txt b/libcxx/include/CMakeLists.txt
index 2ec755236dbaee2..a3d72df61a86dde 100644
--- a/libcxx/include/CMakeLists.txt
+++ b/libcxx/include/CMakeLists.txt
@@ -85,6 +85,11 @@ set(files
   __algorithm/pstl_backends/cpu_backends/thread.h
   __algorithm/pstl_backends/cpu_backends/transform.h
   __algorithm/pstl_backends/cpu_backends/transform_reduce.h
+  __algorithm/pstl_backends/gpu_backend.h
+  __algorithm/pstl_backends/gpu_backends/backend.h
+  __algorithm/pstl_backends/gpu_backends/fill.h
+  __algorithm/pstl_backends/gpu_backends/for_each.h
+  __algorithm/pstl_backends/gpu_backends/omp_offload.h
   __algorithm/pstl_copy.h
   __algorithm/pstl_count.h
   __algorithm/pstl_fill.h
diff --git a/libcxx/include/__algorithm/pstl_backend.h 
b/libcxx/include/__algorithm/pstl_backend.h
index 93372f019031b63..f051e0ce9be13c3 100644
--- a/libcxx/include/__algorithm/pstl_backend.h
+++ b/libcxx/include/__algorithm/pstl_backend.h
@@ -10,6 +10,7 @@
 #define _LIBCPP___ALGORITHM_PSTL_BACKEND_H
 
 #include <__algorithm/pstl_backends/cpu_backend.h>
+#include <__algorithm/pstl_backends/gpu_backend.h>
 #include <__config>
 #include 
 
@@ -179,10 +180,17 @@ struct __select_backend {
   using type = __cpu_backend_tag;
 };
 
+#   if defined(_LIBCPP_PSTL_GPU_BACKEND_OMP_OFFLOAD)
+template <>
+struct __select_backend {
+  using type = __gpu_backend_tag;
+};
+#   else
 template <>
 struct __select_backend {
   using type = __cpu_backend_tag;
 };
+#   endif
 
 #  else
 
diff --git a/libcxx/include/__algorithm/pstl_backends/gpu_backend.h 
b/libcxx/include/__algorithm/pstl_backends/gpu_backend.h
new file mode 100644
index 000..46a85f77b5deb99
--- /dev/null
+++ b/libcxx/include/__algorithm/pstl_backends/gpu_backend.h
@@ -0,0 +1,21 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM 
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--

[clang-tools-extra] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/66968

>From b35340e47de896c9933c54ce617538c46cf01488 Mon Sep 17 00:00:00 2001
From: AntonRydahl 
Date: Wed, 20 Sep 2023 17:06:10 -0700
Subject: [PATCH 1/7] Adding OpenMP Offloading Backend for C++ Parallel
 Algorithms

---
 libcxx/CMakeLists.txt | 14 +++
 libcxx/include/CMakeLists.txt |  5 +
 libcxx/include/__algorithm/pstl_backend.h |  8 ++
 .../__algorithm/pstl_backends/gpu_backend.h   | 21 +
 .../pstl_backends/gpu_backends/backend.h  | 33 +++
 .../pstl_backends/gpu_backends/fill.h | 59 
 .../pstl_backends/gpu_backends/for_each.h | 59 
 .../pstl_backends/gpu_backends/omp_offload.h  | 91 +++
 libcxx/include/__config_site.in   |  1 +
 9 files changed, 291 insertions(+)
 create mode 100644 libcxx/include/__algorithm/pstl_backends/gpu_backend.h
 create mode 100644 
libcxx/include/__algorithm/pstl_backends/gpu_backends/backend.h
 create mode 100644 libcxx/include/__algorithm/pstl_backends/gpu_backends/fill.h
 create mode 100644 
libcxx/include/__algorithm/pstl_backends/gpu_backends/for_each.h
 create mode 100644 
libcxx/include/__algorithm/pstl_backends/gpu_backends/omp_offload.h

diff --git a/libcxx/CMakeLists.txt b/libcxx/CMakeLists.txt
index bb2898b799bcef9..43d2a448de79584 100644
--- a/libcxx/CMakeLists.txt
+++ b/libcxx/CMakeLists.txt
@@ -290,6 +290,8 @@ option(LIBCXX_HAS_WIN32_THREAD_API "Ignore auto-detection 
and force use of win32
 option(LIBCXX_HAS_EXTERNAL_THREAD_API
   "Build libc++ with an externalized threading API.
This option may only be set to ON when LIBCXX_ENABLE_THREADS=ON." OFF)
+option(LIBCXX_ENABLE_GPU_OFFLOAD 
+  "Build libc++ with support for GPU offload" OFF)
 
 if (LIBCXX_ENABLE_THREADS)
   set(LIBCXX_PSTL_CPU_BACKEND "std_thread" CACHE STRING "Which PSTL CPU 
backend to use")
@@ -297,6 +299,14 @@ else()
   set(LIBCXX_PSTL_CPU_BACKEND "serial" CACHE STRING "Which PSTL CPU backend to 
use")
 endif()
 
+if (NOT DEFINED LIBCXX_PSTL_GPU_BACKEND)
+  if (${LIBCXX_ENABLE_GPU_OFFLOAD})
+set(LIBCXX_PSTL_GPU_BACKEND "omp_offload" CACHE STRING "Which PSTL GPU 
backend to use")
+  else()
+set(LIBCXX_PSTL_GPU_BACKEND "none" CACHE STRING "Which PSTL GPU backend to 
use")
+  endif()
+endif()
+
 # Misc options 
 # FIXME: Turn -pedantic back ON. It is currently off because it warns
 # about #include_next which is used everywhere.
@@ -809,6 +819,10 @@ else()
Valid backends are: serial, std_thread and libdispatch")
 endif()
 
+if (LIBCXX_PSTL_GPU_BACKEND STREQUAL "omp_offload")
+  config_define(1 _LIBCPP_PSTL_GPU_BACKEND_OMP_OFFLOAD)
+endif()
+
 if (LIBCXX_ABI_DEFINES)
   set(abi_defines)
   foreach (abi_define ${LIBCXX_ABI_DEFINES})
diff --git a/libcxx/include/CMakeLists.txt b/libcxx/include/CMakeLists.txt
index 2ec755236dbaee2..a3d72df61a86dde 100644
--- a/libcxx/include/CMakeLists.txt
+++ b/libcxx/include/CMakeLists.txt
@@ -85,6 +85,11 @@ set(files
   __algorithm/pstl_backends/cpu_backends/thread.h
   __algorithm/pstl_backends/cpu_backends/transform.h
   __algorithm/pstl_backends/cpu_backends/transform_reduce.h
+  __algorithm/pstl_backends/gpu_backend.h
+  __algorithm/pstl_backends/gpu_backends/backend.h
+  __algorithm/pstl_backends/gpu_backends/fill.h
+  __algorithm/pstl_backends/gpu_backends/for_each.h
+  __algorithm/pstl_backends/gpu_backends/omp_offload.h
   __algorithm/pstl_copy.h
   __algorithm/pstl_count.h
   __algorithm/pstl_fill.h
diff --git a/libcxx/include/__algorithm/pstl_backend.h 
b/libcxx/include/__algorithm/pstl_backend.h
index 93372f019031b63..f051e0ce9be13c3 100644
--- a/libcxx/include/__algorithm/pstl_backend.h
+++ b/libcxx/include/__algorithm/pstl_backend.h
@@ -10,6 +10,7 @@
 #define _LIBCPP___ALGORITHM_PSTL_BACKEND_H
 
 #include <__algorithm/pstl_backends/cpu_backend.h>
+#include <__algorithm/pstl_backends/gpu_backend.h>
 #include <__config>
 #include 
 
@@ -179,10 +180,17 @@ struct __select_backend {
   using type = __cpu_backend_tag;
 };
 
+#   if defined(_LIBCPP_PSTL_GPU_BACKEND_OMP_OFFLOAD)
+template <>
+struct __select_backend {
+  using type = __gpu_backend_tag;
+};
+#   else
 template <>
 struct __select_backend {
   using type = __cpu_backend_tag;
 };
+#   endif
 
 #  else
 
diff --git a/libcxx/include/__algorithm/pstl_backends/gpu_backend.h 
b/libcxx/include/__algorithm/pstl_backends/gpu_backend.h
new file mode 100644
index 000..46a85f77b5deb99
--- /dev/null
+++ b/libcxx/include/__algorithm/pstl_backends/gpu_backend.h
@@ -0,0 +1,21 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM 
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--

[clang-tools-extra] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/66968

>From b35340e47de896c9933c54ce617538c46cf01488 Mon Sep 17 00:00:00 2001
From: AntonRydahl 
Date: Wed, 20 Sep 2023 17:06:10 -0700
Subject: [PATCH 1/8] Adding OpenMP Offloading Backend for C++ Parallel
 Algorithms

---
 libcxx/CMakeLists.txt | 14 +++
 libcxx/include/CMakeLists.txt |  5 +
 libcxx/include/__algorithm/pstl_backend.h |  8 ++
 .../__algorithm/pstl_backends/gpu_backend.h   | 21 +
 .../pstl_backends/gpu_backends/backend.h  | 33 +++
 .../pstl_backends/gpu_backends/fill.h | 59 
 .../pstl_backends/gpu_backends/for_each.h | 59 
 .../pstl_backends/gpu_backends/omp_offload.h  | 91 +++
 libcxx/include/__config_site.in   |  1 +
 9 files changed, 291 insertions(+)
 create mode 100644 libcxx/include/__algorithm/pstl_backends/gpu_backend.h
 create mode 100644 
libcxx/include/__algorithm/pstl_backends/gpu_backends/backend.h
 create mode 100644 libcxx/include/__algorithm/pstl_backends/gpu_backends/fill.h
 create mode 100644 
libcxx/include/__algorithm/pstl_backends/gpu_backends/for_each.h
 create mode 100644 
libcxx/include/__algorithm/pstl_backends/gpu_backends/omp_offload.h

diff --git a/libcxx/CMakeLists.txt b/libcxx/CMakeLists.txt
index bb2898b799bcef9..43d2a448de79584 100644
--- a/libcxx/CMakeLists.txt
+++ b/libcxx/CMakeLists.txt
@@ -290,6 +290,8 @@ option(LIBCXX_HAS_WIN32_THREAD_API "Ignore auto-detection 
and force use of win32
 option(LIBCXX_HAS_EXTERNAL_THREAD_API
   "Build libc++ with an externalized threading API.
This option may only be set to ON when LIBCXX_ENABLE_THREADS=ON." OFF)
+option(LIBCXX_ENABLE_GPU_OFFLOAD 
+  "Build libc++ with support for GPU offload" OFF)
 
 if (LIBCXX_ENABLE_THREADS)
   set(LIBCXX_PSTL_CPU_BACKEND "std_thread" CACHE STRING "Which PSTL CPU 
backend to use")
@@ -297,6 +299,14 @@ else()
   set(LIBCXX_PSTL_CPU_BACKEND "serial" CACHE STRING "Which PSTL CPU backend to 
use")
 endif()
 
+if (NOT DEFINED LIBCXX_PSTL_GPU_BACKEND)
+  if (${LIBCXX_ENABLE_GPU_OFFLOAD})
+set(LIBCXX_PSTL_GPU_BACKEND "omp_offload" CACHE STRING "Which PSTL GPU 
backend to use")
+  else()
+set(LIBCXX_PSTL_GPU_BACKEND "none" CACHE STRING "Which PSTL GPU backend to 
use")
+  endif()
+endif()
+
 # Misc options 
 # FIXME: Turn -pedantic back ON. It is currently off because it warns
 # about #include_next which is used everywhere.
@@ -809,6 +819,10 @@ else()
Valid backends are: serial, std_thread and libdispatch")
 endif()
 
+if (LIBCXX_PSTL_GPU_BACKEND STREQUAL "omp_offload")
+  config_define(1 _LIBCPP_PSTL_GPU_BACKEND_OMP_OFFLOAD)
+endif()
+
 if (LIBCXX_ABI_DEFINES)
   set(abi_defines)
   foreach (abi_define ${LIBCXX_ABI_DEFINES})
diff --git a/libcxx/include/CMakeLists.txt b/libcxx/include/CMakeLists.txt
index 2ec755236dbaee2..a3d72df61a86dde 100644
--- a/libcxx/include/CMakeLists.txt
+++ b/libcxx/include/CMakeLists.txt
@@ -85,6 +85,11 @@ set(files
   __algorithm/pstl_backends/cpu_backends/thread.h
   __algorithm/pstl_backends/cpu_backends/transform.h
   __algorithm/pstl_backends/cpu_backends/transform_reduce.h
+  __algorithm/pstl_backends/gpu_backend.h
+  __algorithm/pstl_backends/gpu_backends/backend.h
+  __algorithm/pstl_backends/gpu_backends/fill.h
+  __algorithm/pstl_backends/gpu_backends/for_each.h
+  __algorithm/pstl_backends/gpu_backends/omp_offload.h
   __algorithm/pstl_copy.h
   __algorithm/pstl_count.h
   __algorithm/pstl_fill.h
diff --git a/libcxx/include/__algorithm/pstl_backend.h 
b/libcxx/include/__algorithm/pstl_backend.h
index 93372f019031b63..f051e0ce9be13c3 100644
--- a/libcxx/include/__algorithm/pstl_backend.h
+++ b/libcxx/include/__algorithm/pstl_backend.h
@@ -10,6 +10,7 @@
 #define _LIBCPP___ALGORITHM_PSTL_BACKEND_H
 
 #include <__algorithm/pstl_backends/cpu_backend.h>
+#include <__algorithm/pstl_backends/gpu_backend.h>
 #include <__config>
 #include 
 
@@ -179,10 +180,17 @@ struct __select_backend {
   using type = __cpu_backend_tag;
 };
 
+#   if defined(_LIBCPP_PSTL_GPU_BACKEND_OMP_OFFLOAD)
+template <>
+struct __select_backend {
+  using type = __gpu_backend_tag;
+};
+#   else
 template <>
 struct __select_backend {
   using type = __cpu_backend_tag;
 };
+#   endif
 
 #  else
 
diff --git a/libcxx/include/__algorithm/pstl_backends/gpu_backend.h 
b/libcxx/include/__algorithm/pstl_backends/gpu_backend.h
new file mode 100644
index 000..46a85f77b5deb99
--- /dev/null
+++ b/libcxx/include/__algorithm/pstl_backends/gpu_backend.h
@@ -0,0 +1,21 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM 
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--

[clang-tools-extra] [libc][libm][GPU] Add missing vendor entrypoints to the GPU version of `libm` (PR #66034)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/66034

>From 7a357059d3ca78afcb02e33a6afce225029a3836 Mon Sep 17 00:00:00 2001
From: antonrydahl 
Date: Mon, 11 Sep 2023 17:06:41 -0700
Subject: [PATCH 1/3] [libc][libm][GPU] Add missing vendor entrypoints to the
 GPU version of `libm`

There are a number of entry-points in `libm` for which no target-agnostic 
implementations exist and the built-ins do not lower correctly on either AMDGPU 
or NVPTX targets. This patch adds the vendor entry-points as a temporary 
solution.
---
 libc/config/gpu/entrypoints.txt   |  37 -
 libc/src/math/CMakeLists.txt  |  15 ++
 libc/src/math/acos.h  |  18 +++
 libc/src/math/acosh.h |  18 +++
 libc/src/math/asin.h  |  18 +++
 libc/src/math/asinh.h |  18 +++
 libc/src/math/atan.h  |  18 +++
 libc/src/math/atan2.h |  18 +++
 libc/src/math/atan2f.h|  18 +++
 libc/src/math/atanh.h |  18 +++
 libc/src/math/erf.h   |  18 +++
 libc/src/math/expm1.h |  18 +++
 libc/src/math/gpu/vendor/CMakeLists.txt   | 143 ++
 libc/src/math/gpu/vendor/acos.cpp |  18 +++
 libc/src/math/gpu/vendor/acosh.cpp|  18 +++
 libc/src/math/gpu/vendor/amdgpu/amdgpu.h  |  25 +++
 .../src/math/gpu/vendor/amdgpu/declarations.h |  34 -
 libc/src/math/gpu/vendor/asin.cpp |  18 +++
 libc/src/math/gpu/vendor/asinh.cpp|  18 +++
 libc/src/math/gpu/vendor/atan.cpp |  18 +++
 libc/src/math/gpu/vendor/atan2.cpp|  20 +++
 libc/src/math/gpu/vendor/atan2f.cpp   |  21 +++
 libc/src/math/gpu/vendor/atanh.cpp|  18 +++
 libc/src/math/gpu/vendor/erf.cpp  |  18 +++
 libc/src/math/gpu/vendor/erff.cpp |  18 +++
 libc/src/math/gpu/vendor/exp.cpp  |  18 +++
 libc/src/math/gpu/vendor/exp10.cpp|  18 +++
 libc/src/math/gpu/vendor/exp2.cpp |  18 +++
 libc/src/math/gpu/vendor/expm1.cpp|  18 +++
 libc/src/math/gpu/vendor/log.cpp  |  18 +++
 libc/src/math/gpu/vendor/log10.cpp|  18 +++
 libc/src/math/gpu/vendor/log10f.cpp   |  19 +++
 libc/src/math/gpu/vendor/log1p.cpp|  18 +++
 libc/src/math/gpu/vendor/log1pf.cpp   |  18 +++
 libc/src/math/gpu/vendor/log2.cpp |  18 +++
 libc/src/math/gpu/vendor/log2f.cpp|  18 +++
 libc/src/math/gpu/vendor/logb.cpp |  18 +++
 libc/src/math/gpu/vendor/logbf.cpp|  19 +++
 libc/src/math/gpu/vendor/logf.cpp |  18 +++
 libc/src/math/gpu/vendor/lrint.cpp|  18 +++
 libc/src/math/gpu/vendor/lrintf.cpp   |  18 +++
 libc/src/math/gpu/vendor/lround.cpp   |  18 +++
 libc/src/math/gpu/vendor/lroundf.cpp  |  20 +++
 libc/src/math/gpu/vendor/nvptx/declarations.h |  25 +++
 libc/src/math/gpu/vendor/nvptx/nvptx.h|  25 +++
 libc/src/math/gpu/vendor/tgamma.cpp   |  18 +++
 libc/src/math/gpu/vendor/tgammaf.cpp  |  18 +++
 libc/src/math/sincos.h|  18 +++
 libc/src/math/tgamma.h|  18 +++
 libc/src/math/tgammaf.h   |  18 +++
 50 files changed, 1081 insertions(+), 6 deletions(-)
 create mode 100644 libc/src/math/acos.h
 create mode 100644 libc/src/math/acosh.h
 create mode 100644 libc/src/math/asin.h
 create mode 100644 libc/src/math/asinh.h
 create mode 100644 libc/src/math/atan.h
 create mode 100644 libc/src/math/atan2.h
 create mode 100644 libc/src/math/atan2f.h
 create mode 100644 libc/src/math/atanh.h
 create mode 100644 libc/src/math/erf.h
 create mode 100644 libc/src/math/expm1.h
 create mode 100644 libc/src/math/gpu/vendor/acos.cpp
 create mode 100644 libc/src/math/gpu/vendor/acosh.cpp
 create mode 100644 libc/src/math/gpu/vendor/asin.cpp
 create mode 100644 libc/src/math/gpu/vendor/asinh.cpp
 create mode 100644 libc/src/math/gpu/vendor/atan.cpp
 create mode 100644 libc/src/math/gpu/vendor/atan2.cpp
 create mode 100644 libc/src/math/gpu/vendor/atan2f.cpp
 create mode 100644 libc/src/math/gpu/vendor/atanh.cpp
 create mode 100644 libc/src/math/gpu/vendor/erf.cpp
 create mode 100644 libc/src/math/gpu/vendor/erff.cpp
 create mode 100644 libc/src/math/gpu/vendor/exp.cpp
 create mode 100644 libc/src/math/gpu/vendor/exp10.cpp
 create mode 100644 libc/src/math/gpu/vendor/exp2.cpp
 create mode 100644 libc/src/math/gpu/vendor/expm1.cpp
 create mode 100644 libc/src/math/gpu/vendor/log.cpp
 create mode 100644 libc/src/math/gpu/vendor/log10.cpp
 create mode 100644 libc/src/math/gpu/vendor/log10f.cpp
 create mode 100644 libc/src/math/gpu/vendor/log1p.cpp
 create mode 100644 libc/src/math/gpu/vendor/log1pf.cpp
 create mode 100644 libc/src/math/gpu/vendor/log2.cpp
 creat

[clang-tools-extra] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

@@ -807,6 +815,11 @@ else()
Valid backends are: serial, std_thread and libdispatch")
 endif()
 
+config_define_if(LIBCXX_ENABLE_GPU_OFFLOAD _LIBCPP_PSTL_GPU_OFFLOAD)

AntonRydahl wrote:

I have tried to restructure the CMake file as sugegsted. However, I had issues 
with making a configuration variable that could be overridden. Do you know what 
type of cache should be used in `set(LIBCXX_PSTL_BACKEND ...)` to allow users 
to override it? 

https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

@@ -807,6 +815,11 @@ else()
Valid backends are: serial, std_thread and libdispatch")
 endif()
 
+config_define_if(LIBCXX_ENABLE_GPU_OFFLOAD _LIBCPP_PSTL_GPU_OFFLOAD)

AntonRydahl wrote:

I have tried to restructure the CMake file as sugegsted. However, I had issues 
with making a configuration variable that could be overridden. Do you know what 
type of cache should be used in `set(LIBCXX_PSTL_BACKEND ...)` to allow users 
to override it? 

https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang-tools-extra] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl edited 
https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

@@ -0,0 +1,120 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM 
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--===//
+
+#ifndef _LIBCPP___ALGORITHM_PSTL_BACKENDS_GPU_BACKENDS_TRANSFORM_REDUCE_H
+#define _LIBCPP___ALGORITHM_PSTL_BACKENDS_GPU_BACKENDS_TRANSFORM_REDUCE_H
+
+#include <__algorithm/pstl_backends/cpu_backends/backend.h>
+#include <__algorithm/pstl_backends/gpu_backends/backend.h>
+#include <__config>
+#include <__functional/operations.h>
+#include <__iterator/concepts.h>
+#include <__iterator/iterator_traits.h>
+#include <__numeric/transform_reduce.h>
+#include <__type_traits/integral_constant.h>
+#include <__type_traits/is_arithmetic.h>
+#include <__type_traits/is_execution_policy.h>
+#include <__type_traits/operation_traits.h>
+#include <__utility/move.h>
+#include <__utility/terminate_on_exception.h>
+#include 
+
+#if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
+#  pragma GCC system_header
+#endif
+
+#if !defined(_LIBCPP_HAS_NO_INCOMPLETE_PSTL) && _LIBCPP_STD_VER >= 17
+
+_LIBCPP_BEGIN_NAMESPACE_STD
+
+template 
+struct _LIBCPP_HIDE_FROM_ABI __is_supported_reduction : std::false_type {};
+
+#  define __PSTL_IS_SUPPORTED_REDUCTION(funname)   
\

AntonRydahl wrote:

To not make this PR even bigger, I have added a refactor to use `__desugars_to 
` here: https://github.com/llvm/llvm-project/pull/68642

I hope that is what you meant. 😄 

https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl edited 
https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang-tools-extra] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

@@ -0,0 +1,120 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM 
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--===//
+
+#ifndef _LIBCPP___ALGORITHM_PSTL_BACKENDS_GPU_BACKENDS_TRANSFORM_REDUCE_H
+#define _LIBCPP___ALGORITHM_PSTL_BACKENDS_GPU_BACKENDS_TRANSFORM_REDUCE_H
+
+#include <__algorithm/pstl_backends/cpu_backends/backend.h>
+#include <__algorithm/pstl_backends/gpu_backends/backend.h>
+#include <__config>
+#include <__functional/operations.h>
+#include <__iterator/concepts.h>
+#include <__iterator/iterator_traits.h>
+#include <__numeric/transform_reduce.h>
+#include <__type_traits/integral_constant.h>
+#include <__type_traits/is_arithmetic.h>
+#include <__type_traits/is_execution_policy.h>
+#include <__type_traits/operation_traits.h>
+#include <__utility/move.h>
+#include <__utility/terminate_on_exception.h>
+#include 
+
+#if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
+#  pragma GCC system_header
+#endif
+
+#if !defined(_LIBCPP_HAS_NO_INCOMPLETE_PSTL) && _LIBCPP_STD_VER >= 17
+
+_LIBCPP_BEGIN_NAMESPACE_STD
+
+template 
+struct _LIBCPP_HIDE_FROM_ABI __is_supported_reduction : std::false_type {};
+
+#  define __PSTL_IS_SUPPORTED_REDUCTION(funname)   
\

AntonRydahl wrote:

To not make this PR even bigger, I have added a refactor to use `__desugars_to 
` here: https://github.com/llvm/llvm-project/pull/68642

I hope that is what you meant. 😄 

https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang-tools-extra] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

@@ -0,0 +1,87 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM 
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--===//
+
+#ifndef _LIBCPP___ALGORITHM_PSTL_BACKENDS_GPU_BACKENDS_TRANSFORM_H
+#define _LIBCPP___ALGORITHM_PSTL_BACKENDS_GPU_BACKENDS_TRANSFORM_H
+
+#include <__algorithm/pstl_backends/cpu_backends/backend.h>
+#include <__algorithm/pstl_backends/gpu_backends/backend.h>
+#include <__algorithm/transform.h>
+#include <__config>
+#include <__iterator/concepts.h>
+#include <__iterator/iterator_traits.h>
+#include <__type_traits/enable_if.h>
+#include <__type_traits/is_execution_policy.h>
+#include <__type_traits/remove_cvref.h>
+#include <__utility/terminate_on_exception.h>
+
+#if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
+#  pragma GCC system_header
+#endif
+
+#if !defined(_LIBCPP_HAS_NO_INCOMPLETE_PSTL) && _LIBCPP_STD_VER >= 17
+
+_LIBCPP_BEGIN_NAMESPACE_STD
+
+template 
+_LIBCPP_HIDE_FROM_ABI _ForwardOutIterator __pstl_transform(
+__gpu_backend_tag,
+_ForwardIterator __first,
+_ForwardIterator __last,
+_ForwardOutIterator __result,
+_UnaryOperation __op) {
+  // The interface for the function switched between C++17 and C++20
+  if constexpr (__is_unsequenced_execution_policy_v<_ExecutionPolicy> &&

AntonRydahl wrote:

I think we could perhaps make OpenMP the backend for both parallel and parallel 
unsequential execution policies. In OpenMP, `#pragma omp target if(cond)` can 
be used to only offload based on a condition. This could be used to execute the 
host fallback code for `std::par`.

https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

@@ -0,0 +1,87 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM 
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--===//
+
+#ifndef _LIBCPP___ALGORITHM_PSTL_BACKENDS_GPU_BACKENDS_TRANSFORM_H
+#define _LIBCPP___ALGORITHM_PSTL_BACKENDS_GPU_BACKENDS_TRANSFORM_H
+
+#include <__algorithm/pstl_backends/cpu_backends/backend.h>
+#include <__algorithm/pstl_backends/gpu_backends/backend.h>
+#include <__algorithm/transform.h>
+#include <__config>
+#include <__iterator/concepts.h>
+#include <__iterator/iterator_traits.h>
+#include <__type_traits/enable_if.h>
+#include <__type_traits/is_execution_policy.h>
+#include <__type_traits/remove_cvref.h>
+#include <__utility/terminate_on_exception.h>
+
+#if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
+#  pragma GCC system_header
+#endif
+
+#if !defined(_LIBCPP_HAS_NO_INCOMPLETE_PSTL) && _LIBCPP_STD_VER >= 17
+
+_LIBCPP_BEGIN_NAMESPACE_STD
+
+template 
+_LIBCPP_HIDE_FROM_ABI _ForwardOutIterator __pstl_transform(
+__gpu_backend_tag,
+_ForwardIterator __first,
+_ForwardIterator __last,
+_ForwardOutIterator __result,
+_UnaryOperation __op) {
+  // The interface for the function switched between C++17 and C++20
+  if constexpr (__is_unsequenced_execution_policy_v<_ExecutionPolicy> &&

AntonRydahl wrote:

I think we could perhaps make OpenMP the backend for both parallel and parallel 
unsequential execution policies. In OpenMP, `#pragma omp target if(cond)` can 
be used to only offload based on a condition. This could be used to execute the 
host fallback code for `std::par`.

https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang-tools-extra] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl edited 
https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl edited 
https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang-tools-extra] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

@@ -0,0 +1,41 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM 
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--===//

AntonRydahl wrote:

I think there is unfortunately a bug in the PSTL logic. I cannot compile 
`libcxx` if I remove the overlay header files for `any_of`, `find_if`, or 
`stable_sort`.

https://github.com/llvm/llvm-project/pull/66968
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/68642

>From f0d93cc6a5cd485c654ab38221691db038bacc7d Mon Sep 17 00:00:00 2001
From: AntonRydahl 
Date: Mon, 9 Oct 2023 15:13:22 -0700
Subject: [PATCH 1/2] Merged __is_trivial_equality_predicate and
 __is_trivial_plus_operation into __desugars_to

---
 libcxx/include/CMakeLists.txt |  1 -
 libcxx/include/__algorithm/comp.h |  7 ++---
 libcxx/include/__algorithm/equal.h| 22 
 .../cpu_backends/transform_reduce.h   | 23 
 libcxx/include/__functional/operations.h  | 15 +--
 .../include/__functional/ranges_operations.h  |  7 ++---
 .../include/__numeric/pstl_transform_reduce.h |  2 +-
 .../include/__type_traits/operation_traits.h  |  4 +--
 .../include/__type_traits/predicate_traits.h  | 26 ---
 libcxx/include/module.modulemap.in|  1 -
 10 files changed, 40 insertions(+), 68 deletions(-)
 delete mode 100644 libcxx/include/__type_traits/predicate_traits.h

diff --git a/libcxx/include/CMakeLists.txt b/libcxx/include/CMakeLists.txt
index 340353f8ebb41c4..7eb09a06ccd482e 100644
--- a/libcxx/include/CMakeLists.txt
+++ b/libcxx/include/CMakeLists.txt
@@ -813,7 +813,6 @@ set(files
   __type_traits/negation.h
   __type_traits/noexcept_move_assign_container.h
   __type_traits/operation_traits.h
-  __type_traits/predicate_traits.h
   __type_traits/promote.h
   __type_traits/rank.h
   __type_traits/remove_all_extents.h
diff --git a/libcxx/include/__algorithm/comp.h 
b/libcxx/include/__algorithm/comp.h
index 9474536615ffb67..0993c37cce36a6b 100644
--- a/libcxx/include/__algorithm/comp.h
+++ b/libcxx/include/__algorithm/comp.h
@@ -10,8 +10,9 @@
 #define _LIBCPP___ALGORITHM_COMP_H
 
 #include <__config>
+#include <__functional/operations.h>
 #include <__type_traits/integral_constant.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
 #  pragma GCC system_header
@@ -26,8 +27,8 @@ struct __equal_to {
   }
 };
 
-template 
-struct __is_trivial_equality_predicate<__equal_to, _Lhs, _Rhs> : true_type {};
+template <>
+struct __desugars_to<__equal_to, std::equal_to<>> : true_type {};
 
 // The definition is required because __less is part of the ABI, but it's empty
 // because all comparisons should be transparent.
diff --git a/libcxx/include/__algorithm/equal.h 
b/libcxx/include/__algorithm/equal.h
index b69aeff92bb9289..35e82da15e4d058 100644
--- a/libcxx/include/__algorithm/equal.h
+++ b/libcxx/include/__algorithm/equal.h
@@ -15,6 +15,7 @@
 #include <__config>
 #include <__functional/identity.h>
 #include <__functional/invoke.h>
+#include <__functional/operations.h>
 #include <__iterator/distance.h>
 #include <__iterator/iterator_traits.h>
 #include <__string/constexpr_c_functions.h>
@@ -23,7 +24,7 @@
 #include <__type_traits/is_constant_evaluated.h>
 #include <__type_traits/is_equality_comparable.h>
 #include <__type_traits/is_volatile.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 #include <__utility/move.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
@@ -41,13 +42,12 @@ _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI 
_LIBCPP_CONSTEXPR_SINCE_CXX20 boo
   return true;
 }
 
-template <
-class _Tp,
-class _Up,
-class _BinaryPredicate,
-__enable_if_t<__is_trivial_equality_predicate<_BinaryPredicate, _Tp, 
_Up>::value && !is_volatile<_Tp>::value &&
-  !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
-  int> = 0>
+template < class _Tp,
+   class _Up,
+   class _BinaryPredicate,
+   __enable_if_t<__desugars_to<_BinaryPredicate, 
std::equal_to<>>::value && !is_volatile<_Tp>::value &&
+ !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
+ int> = 0>
 _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
 __equal_iter_impl(_Tp* __first1, _Tp* __last1, _Up* __first2, 
_BinaryPredicate&) {
   return std::__constexpr_memcmp_equal(__first1, __first2, 
__element_count(__last1 - __first1));
@@ -94,12 +94,12 @@ template ::value && __is_identity<_Proj1>::value &&
+  __enable_if_t<__desugars_to<_Pred, std::equal_to<>>::value && 
__is_identity<_Proj1>::value &&
 __is_identity<_Proj2>::value && 
!is_volatile<_Tp>::value && !is_volatile<_Up>::value &&
 __libcpp_is_trivially_equality_comparable<_Tp, 
_Up>::value,
 int> = 0>
-_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool __equal_impl(
-_Tp* __first1, _Tp* __last1, _Up* __first2, _Up*, _Pred&, _Proj1&, 
_Proj2&) {
+_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
+

[clang-tools-extra] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/68642

>From f0d93cc6a5cd485c654ab38221691db038bacc7d Mon Sep 17 00:00:00 2001
From: AntonRydahl 
Date: Mon, 9 Oct 2023 15:13:22 -0700
Subject: [PATCH 1/2] Merged __is_trivial_equality_predicate and
 __is_trivial_plus_operation into __desugars_to

---
 libcxx/include/CMakeLists.txt |  1 -
 libcxx/include/__algorithm/comp.h |  7 ++---
 libcxx/include/__algorithm/equal.h| 22 
 .../cpu_backends/transform_reduce.h   | 23 
 libcxx/include/__functional/operations.h  | 15 +--
 .../include/__functional/ranges_operations.h  |  7 ++---
 .../include/__numeric/pstl_transform_reduce.h |  2 +-
 .../include/__type_traits/operation_traits.h  |  4 +--
 .../include/__type_traits/predicate_traits.h  | 26 ---
 libcxx/include/module.modulemap.in|  1 -
 10 files changed, 40 insertions(+), 68 deletions(-)
 delete mode 100644 libcxx/include/__type_traits/predicate_traits.h

diff --git a/libcxx/include/CMakeLists.txt b/libcxx/include/CMakeLists.txt
index 340353f8ebb41c4..7eb09a06ccd482e 100644
--- a/libcxx/include/CMakeLists.txt
+++ b/libcxx/include/CMakeLists.txt
@@ -813,7 +813,6 @@ set(files
   __type_traits/negation.h
   __type_traits/noexcept_move_assign_container.h
   __type_traits/operation_traits.h
-  __type_traits/predicate_traits.h
   __type_traits/promote.h
   __type_traits/rank.h
   __type_traits/remove_all_extents.h
diff --git a/libcxx/include/__algorithm/comp.h 
b/libcxx/include/__algorithm/comp.h
index 9474536615ffb67..0993c37cce36a6b 100644
--- a/libcxx/include/__algorithm/comp.h
+++ b/libcxx/include/__algorithm/comp.h
@@ -10,8 +10,9 @@
 #define _LIBCPP___ALGORITHM_COMP_H
 
 #include <__config>
+#include <__functional/operations.h>
 #include <__type_traits/integral_constant.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
 #  pragma GCC system_header
@@ -26,8 +27,8 @@ struct __equal_to {
   }
 };
 
-template 
-struct __is_trivial_equality_predicate<__equal_to, _Lhs, _Rhs> : true_type {};
+template <>
+struct __desugars_to<__equal_to, std::equal_to<>> : true_type {};
 
 // The definition is required because __less is part of the ABI, but it's empty
 // because all comparisons should be transparent.
diff --git a/libcxx/include/__algorithm/equal.h 
b/libcxx/include/__algorithm/equal.h
index b69aeff92bb9289..35e82da15e4d058 100644
--- a/libcxx/include/__algorithm/equal.h
+++ b/libcxx/include/__algorithm/equal.h
@@ -15,6 +15,7 @@
 #include <__config>
 #include <__functional/identity.h>
 #include <__functional/invoke.h>
+#include <__functional/operations.h>
 #include <__iterator/distance.h>
 #include <__iterator/iterator_traits.h>
 #include <__string/constexpr_c_functions.h>
@@ -23,7 +24,7 @@
 #include <__type_traits/is_constant_evaluated.h>
 #include <__type_traits/is_equality_comparable.h>
 #include <__type_traits/is_volatile.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 #include <__utility/move.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
@@ -41,13 +42,12 @@ _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI 
_LIBCPP_CONSTEXPR_SINCE_CXX20 boo
   return true;
 }
 
-template <
-class _Tp,
-class _Up,
-class _BinaryPredicate,
-__enable_if_t<__is_trivial_equality_predicate<_BinaryPredicate, _Tp, 
_Up>::value && !is_volatile<_Tp>::value &&
-  !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
-  int> = 0>
+template < class _Tp,
+   class _Up,
+   class _BinaryPredicate,
+   __enable_if_t<__desugars_to<_BinaryPredicate, 
std::equal_to<>>::value && !is_volatile<_Tp>::value &&
+ !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
+ int> = 0>
 _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
 __equal_iter_impl(_Tp* __first1, _Tp* __last1, _Up* __first2, 
_BinaryPredicate&) {
   return std::__constexpr_memcmp_equal(__first1, __first2, 
__element_count(__last1 - __first1));
@@ -94,12 +94,12 @@ template ::value && __is_identity<_Proj1>::value &&
+  __enable_if_t<__desugars_to<_Pred, std::equal_to<>>::value && 
__is_identity<_Proj1>::value &&
 __is_identity<_Proj2>::value && 
!is_volatile<_Tp>::value && !is_volatile<_Up>::value &&
 __libcpp_is_trivially_equality_comparable<_Tp, 
_Up>::value,
 int> = 0>
-_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool __equal_impl(
-_Tp* __first1, _Tp* __last1, _Up* __first2, _Up*, _Pred&, _Proj1&, 
_Proj2&) {
+_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
+

[clang] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

@@ -0,0 +1,33 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM 
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--===//
+
+#ifndef _LIBCPP___ALGORITHM_PSTL_BACKENDS_OPENMP_BACKEND_ANY_OF_H
+#define _LIBCPP___ALGORITHM_PSTL_BACKENDS_OPENMP_BACKEND_ANY_OF_H
+
+#include <__algorithm/any_of.h>
+#include <__algorithm/pstl_backends/cpu_backends/backend.h>
+#include <__algorithm/pstl_backends/openmp/backend.h>
+#include <__config>
+#include <__type_traits/is_execution_policy.h>
+
+#if !defined(_LIBCPP_HAS_NO_INCOMPLETE_PSTL) && _LIBCPP_STD_VER >= 17
+
+_LIBCPP_BEGIN_NAMESPACE_STD
+
+template 
+_LIBCPP_HIDE_FROM_ABI bool
+__pstl_any_of(__omp_backend_tag, _ForwardIterator __first, _ForwardIterator 
__last, _Predicate __pred) {

AntonRydahl wrote:

Here is the full error for one of the tests that fail if I do not make a 
placeholder for any_of:
```
 TEST 'llvm-libc++-shared.cfg.in :: 
std/algorithms/alg.modifying.operations/alg.fill/pstl.fill_n.pass.cpp' FAILED 

Exit Code: 1

Command Output (stdout):
--
# COMPILED WITH
/dev/shm/rydahl1/LLVM/build/./bin/clang++ 
/g/g92/rydahl1/LLVM_FORK/llvm-project/libcxx/test/std/algorithms/alg.modifying.operations/alg.fill/pstl.fill_n.pass.cpp
 -pthread --target=x86_64-unknown-linux-gnu -nostdinc++ -I 
/dev/shm/rydahl1/LLVM/build/include/c++/v1 -I 
/dev/shm/rydahl1/LLVM/build/include/x86_64-unknown-linux-gnu/c++/v1 -I 
/g/g92/rydahl1/LLVM_FORK/llvm-project/libcxx/test/support -std=c++26 -Werror 
-Wall -Wctad-maybe-unsupported -Wextra -Wshadow -Wundef -Wunused-template 
-Wno-unused-command-line-argument -Wno-attributes -Wno-pessimizing-move 
-Wno-noexcept-type -Wno-atomic-alignment -Wno-reserved-module-identifier 
-Wdeprecated-copy -Wdeprecated-copy-dtor -Wno-user-defined-literals 
-Wno-tautological-compare -Wsign-compare -Wunused-variable -Wunused-parameter 
-Wunreachable-code -Wno-unused-local-typedef -Wno-local-type-template-args 
-Wno-c++11-extensions -Wno-unknown-pragmas -Wno-pass-failed 
-Wno-mismatched-new-delete -Wno-redundant-move -Wno-self-move 
-D_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER -D_LIBCPP_ENABLE_EXPERIMENTAL 
-Werror=thread-safety -Wuser-defined-warnings  -lc++experimental -nostdlib++ -L 
/dev/shm/rydahl1/LLVM/build/./lib/x86_64-unknown-linux-gnu 
-Wl,-rpath,/dev/shm/rydahl1/LLVM/build/./lib/x86_64-unknown-linux-gnu -lc++ -o 
/dev/shm/rydahl1/LLVM/build/runtimes/runtimes-bins/test/std/algorithms/alg.modifying.operations/alg.fill/Output/pstl.fill_n.pass.cpp.dir/t.tmp.exe
# executed command: /dev/shm/rydahl1/LLVM/build/./bin/clang++ 
/g/g92/rydahl1/LLVM_FORK/llvm-project/libcxx/test/std/algorithms/alg.modifying.operations/alg.fill/pstl.fill_n.pass.cpp
 -pthread --target=x86_64-unknown-linux-gnu -nostdinc++ -I 
/dev/shm/rydahl1/LLVM/build/include/c++/v1 -I 
/dev/shm/rydahl1/LLVM/build/include/x86_64-unknown-linux-gnu/c++/v1 -I 
/g/g92/rydahl1/LLVM_FORK/llvm-project/libcxx/test/support -std=c++26 -Werror 
-Wall -Wctad-maybe-unsupported -Wextra -Wshadow -Wundef -Wunused-template 
-Wno-unused-command-line-argument -Wno-attributes -Wno-pessimizing-move 
-Wno-noexcept-type -Wno-atomic-alignment -Wno-reserved-module-identifier 
-Wdeprecated-copy -Wdeprecated-copy-dtor -Wno-user-defined-literals 
-Wno-tautological-compare -Wsign-compare -Wunused-variable -Wunused-parameter 
-Wunreachable-code -Wno-unused-local-typedef -Wno-local-type-template-args 
-Wno-c++11-extensions -Wno-unknown-pragmas -Wno-pass-failed 
-Wno-mismatched-new-delete -Wno-redundant-move -Wno-self-move 
-D_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER -D_LIBCPP_ENABLE_EXPERIMENTAL 
-Werror=thread-safety -Wuser-defined-warnings -lc++experimental -nostdlib++ -L 
/dev/shm/rydahl1/LLVM/build/./lib/x86_64-unknown-linux-gnu 
-Wl,-rpath,/dev/shm/rydahl1/LLVM/build/./lib/x86_64-unknown-linux-gnu -lc++ -o 
/dev/shm/rydahl1/LLVM/build/runtimes/runtimes-bins/test/std/algorithms/alg.modifying.operations/alg.fill/Output/pstl.fill_n.pass.cpp.dir/t.tmp.exe
# .---command stderr
# | In file included from 
/g/g92/rydahl1/LLVM_FORK/llvm-project/libcxx/test/std/algorithms/alg.modifying.operations/alg.fill/pstl.fill_n.pass.cpp:19:
# | In file included from 
/dev/shm/rydahl1/LLVM/build/include/c++/v1/algorithm:1827:
# | In file included from 
/dev/shm/rydahl1/LLVM/build/include/c++/v1/__algorithm/pstl_any_all_none_of.h:12:
# | In file included from 
/dev/shm/rydahl1/LLVM/build/include/c++/v1/__algorithm/pstl_find.h:14:
# | In file included from 
/dev/shm/rydahl1/LLVM/build/include/c++/v1/__algorithm/pstl_backend.h:12:
# | In file included from 
/dev/shm/rydahl1/LLVM/build/include/c++/v1/__algorithm/pstl_backends/cpu_backend.h:58:
# | In file included from 
/dev/shm/rydahl1/LLVM/build/inclu

[clang-tools-extra] Adding Separate OpenMP Offloading Backend to `libcxx/include/__algorithm/pstl_backends` (PR #66968)

[Anton Rydahl via cfe-commits]

@@ -0,0 +1,33 @@
+//===--===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM 
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===--===//
+
+#ifndef _LIBCPP___ALGORITHM_PSTL_BACKENDS_OPENMP_BACKEND_ANY_OF_H
+#define _LIBCPP___ALGORITHM_PSTL_BACKENDS_OPENMP_BACKEND_ANY_OF_H
+
+#include <__algorithm/any_of.h>
+#include <__algorithm/pstl_backends/cpu_backends/backend.h>
+#include <__algorithm/pstl_backends/openmp/backend.h>
+#include <__config>
+#include <__type_traits/is_execution_policy.h>
+
+#if !defined(_LIBCPP_HAS_NO_INCOMPLETE_PSTL) && _LIBCPP_STD_VER >= 17
+
+_LIBCPP_BEGIN_NAMESPACE_STD
+
+template 
+_LIBCPP_HIDE_FROM_ABI bool
+__pstl_any_of(__omp_backend_tag, _ForwardIterator __first, _ForwardIterator 
__last, _Predicate __pred) {

AntonRydahl wrote:

Here is the full error for one of the tests that fail if I do not make a 
placeholder for any_of:
```
 TEST 'llvm-libc++-shared.cfg.in :: 
std/algorithms/alg.modifying.operations/alg.fill/pstl.fill_n.pass.cpp' FAILED 

Exit Code: 1

Command Output (stdout):
--
# COMPILED WITH
/dev/shm/rydahl1/LLVM/build/./bin/clang++ 
/g/g92/rydahl1/LLVM_FORK/llvm-project/libcxx/test/std/algorithms/alg.modifying.operations/alg.fill/pstl.fill_n.pass.cpp
 -pthread --target=x86_64-unknown-linux-gnu -nostdinc++ -I 
/dev/shm/rydahl1/LLVM/build/include/c++/v1 -I 
/dev/shm/rydahl1/LLVM/build/include/x86_64-unknown-linux-gnu/c++/v1 -I 
/g/g92/rydahl1/LLVM_FORK/llvm-project/libcxx/test/support -std=c++26 -Werror 
-Wall -Wctad-maybe-unsupported -Wextra -Wshadow -Wundef -Wunused-template 
-Wno-unused-command-line-argument -Wno-attributes -Wno-pessimizing-move 
-Wno-noexcept-type -Wno-atomic-alignment -Wno-reserved-module-identifier 
-Wdeprecated-copy -Wdeprecated-copy-dtor -Wno-user-defined-literals 
-Wno-tautological-compare -Wsign-compare -Wunused-variable -Wunused-parameter 
-Wunreachable-code -Wno-unused-local-typedef -Wno-local-type-template-args 
-Wno-c++11-extensions -Wno-unknown-pragmas -Wno-pass-failed 
-Wno-mismatched-new-delete -Wno-redundant-move -Wno-self-move 
-D_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER -D_LIBCPP_ENABLE_EXPERIMENTAL 
-Werror=thread-safety -Wuser-defined-warnings  -lc++experimental -nostdlib++ -L 
/dev/shm/rydahl1/LLVM/build/./lib/x86_64-unknown-linux-gnu 
-Wl,-rpath,/dev/shm/rydahl1/LLVM/build/./lib/x86_64-unknown-linux-gnu -lc++ -o 
/dev/shm/rydahl1/LLVM/build/runtimes/runtimes-bins/test/std/algorithms/alg.modifying.operations/alg.fill/Output/pstl.fill_n.pass.cpp.dir/t.tmp.exe
# executed command: /dev/shm/rydahl1/LLVM/build/./bin/clang++ 
/g/g92/rydahl1/LLVM_FORK/llvm-project/libcxx/test/std/algorithms/alg.modifying.operations/alg.fill/pstl.fill_n.pass.cpp
 -pthread --target=x86_64-unknown-linux-gnu -nostdinc++ -I 
/dev/shm/rydahl1/LLVM/build/include/c++/v1 -I 
/dev/shm/rydahl1/LLVM/build/include/x86_64-unknown-linux-gnu/c++/v1 -I 
/g/g92/rydahl1/LLVM_FORK/llvm-project/libcxx/test/support -std=c++26 -Werror 
-Wall -Wctad-maybe-unsupported -Wextra -Wshadow -Wundef -Wunused-template 
-Wno-unused-command-line-argument -Wno-attributes -Wno-pessimizing-move 
-Wno-noexcept-type -Wno-atomic-alignment -Wno-reserved-module-identifier 
-Wdeprecated-copy -Wdeprecated-copy-dtor -Wno-user-defined-literals 
-Wno-tautological-compare -Wsign-compare -Wunused-variable -Wunused-parameter 
-Wunreachable-code -Wno-unused-local-typedef -Wno-local-type-template-args 
-Wno-c++11-extensions -Wno-unknown-pragmas -Wno-pass-failed 
-Wno-mismatched-new-delete -Wno-redundant-move -Wno-self-move 
-D_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER -D_LIBCPP_ENABLE_EXPERIMENTAL 
-Werror=thread-safety -Wuser-defined-warnings -lc++experimental -nostdlib++ -L 
/dev/shm/rydahl1/LLVM/build/./lib/x86_64-unknown-linux-gnu 
-Wl,-rpath,/dev/shm/rydahl1/LLVM/build/./lib/x86_64-unknown-linux-gnu -lc++ -o 
/dev/shm/rydahl1/LLVM/build/runtimes/runtimes-bins/test/std/algorithms/alg.modifying.operations/alg.fill/Output/pstl.fill_n.pass.cpp.dir/t.tmp.exe
# .---command stderr
# | In file included from 
/g/g92/rydahl1/LLVM_FORK/llvm-project/libcxx/test/std/algorithms/alg.modifying.operations/alg.fill/pstl.fill_n.pass.cpp:19:
# | In file included from 
/dev/shm/rydahl1/LLVM/build/include/c++/v1/algorithm:1827:
# | In file included from 
/dev/shm/rydahl1/LLVM/build/include/c++/v1/__algorithm/pstl_any_all_none_of.h:12:
# | In file included from 
/dev/shm/rydahl1/LLVM/build/include/c++/v1/__algorithm/pstl_find.h:14:
# | In file included from 
/dev/shm/rydahl1/LLVM/build/include/c++/v1/__algorithm/pstl_backend.h:12:
# | In file included from 
/dev/shm/rydahl1/LLVM/build/include/c++/v1/__algorithm/pstl_backends/cpu_backend.h:58:
# | In file included from 
/dev/shm/rydahl1/LLVM/build/inclu

[clang] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/68642

>From f0d93cc6a5cd485c654ab38221691db038bacc7d Mon Sep 17 00:00:00 2001
From: AntonRydahl 
Date: Mon, 9 Oct 2023 15:13:22 -0700
Subject: [PATCH 1/3] Merged __is_trivial_equality_predicate and
 __is_trivial_plus_operation into __desugars_to

---
 libcxx/include/CMakeLists.txt |  1 -
 libcxx/include/__algorithm/comp.h |  7 ++---
 libcxx/include/__algorithm/equal.h| 22 
 .../cpu_backends/transform_reduce.h   | 23 
 libcxx/include/__functional/operations.h  | 15 +--
 .../include/__functional/ranges_operations.h  |  7 ++---
 .../include/__numeric/pstl_transform_reduce.h |  2 +-
 .../include/__type_traits/operation_traits.h  |  4 +--
 .../include/__type_traits/predicate_traits.h  | 26 ---
 libcxx/include/module.modulemap.in|  1 -
 10 files changed, 40 insertions(+), 68 deletions(-)
 delete mode 100644 libcxx/include/__type_traits/predicate_traits.h

diff --git a/libcxx/include/CMakeLists.txt b/libcxx/include/CMakeLists.txt
index 340353f8ebb41c4..7eb09a06ccd482e 100644
--- a/libcxx/include/CMakeLists.txt
+++ b/libcxx/include/CMakeLists.txt
@@ -813,7 +813,6 @@ set(files
   __type_traits/negation.h
   __type_traits/noexcept_move_assign_container.h
   __type_traits/operation_traits.h
-  __type_traits/predicate_traits.h
   __type_traits/promote.h
   __type_traits/rank.h
   __type_traits/remove_all_extents.h
diff --git a/libcxx/include/__algorithm/comp.h 
b/libcxx/include/__algorithm/comp.h
index 9474536615ffb67..0993c37cce36a6b 100644
--- a/libcxx/include/__algorithm/comp.h
+++ b/libcxx/include/__algorithm/comp.h
@@ -10,8 +10,9 @@
 #define _LIBCPP___ALGORITHM_COMP_H
 
 #include <__config>
+#include <__functional/operations.h>
 #include <__type_traits/integral_constant.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
 #  pragma GCC system_header
@@ -26,8 +27,8 @@ struct __equal_to {
   }
 };
 
-template 
-struct __is_trivial_equality_predicate<__equal_to, _Lhs, _Rhs> : true_type {};
+template <>
+struct __desugars_to<__equal_to, std::equal_to<>> : true_type {};
 
 // The definition is required because __less is part of the ABI, but it's empty
 // because all comparisons should be transparent.
diff --git a/libcxx/include/__algorithm/equal.h 
b/libcxx/include/__algorithm/equal.h
index b69aeff92bb9289..35e82da15e4d058 100644
--- a/libcxx/include/__algorithm/equal.h
+++ b/libcxx/include/__algorithm/equal.h
@@ -15,6 +15,7 @@
 #include <__config>
 #include <__functional/identity.h>
 #include <__functional/invoke.h>
+#include <__functional/operations.h>
 #include <__iterator/distance.h>
 #include <__iterator/iterator_traits.h>
 #include <__string/constexpr_c_functions.h>
@@ -23,7 +24,7 @@
 #include <__type_traits/is_constant_evaluated.h>
 #include <__type_traits/is_equality_comparable.h>
 #include <__type_traits/is_volatile.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 #include <__utility/move.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
@@ -41,13 +42,12 @@ _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI 
_LIBCPP_CONSTEXPR_SINCE_CXX20 boo
   return true;
 }
 
-template <
-class _Tp,
-class _Up,
-class _BinaryPredicate,
-__enable_if_t<__is_trivial_equality_predicate<_BinaryPredicate, _Tp, 
_Up>::value && !is_volatile<_Tp>::value &&
-  !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
-  int> = 0>
+template < class _Tp,
+   class _Up,
+   class _BinaryPredicate,
+   __enable_if_t<__desugars_to<_BinaryPredicate, 
std::equal_to<>>::value && !is_volatile<_Tp>::value &&
+ !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
+ int> = 0>
 _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
 __equal_iter_impl(_Tp* __first1, _Tp* __last1, _Up* __first2, 
_BinaryPredicate&) {
   return std::__constexpr_memcmp_equal(__first1, __first2, 
__element_count(__last1 - __first1));
@@ -94,12 +94,12 @@ template ::value && __is_identity<_Proj1>::value &&
+  __enable_if_t<__desugars_to<_Pred, std::equal_to<>>::value && 
__is_identity<_Proj1>::value &&
 __is_identity<_Proj2>::value && 
!is_volatile<_Tp>::value && !is_volatile<_Up>::value &&
 __libcpp_is_trivially_equality_comparable<_Tp, 
_Up>::value,
 int> = 0>
-_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool __equal_impl(
-_Tp* __first1, _Tp* __last1, _Up* __first2, _Up*, _Pred&, _Proj1&, 
_Proj2&) {
+_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
+

[clang-tools-extra] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/68642

>From f0d93cc6a5cd485c654ab38221691db038bacc7d Mon Sep 17 00:00:00 2001
From: AntonRydahl 
Date: Mon, 9 Oct 2023 15:13:22 -0700
Subject: [PATCH 1/3] Merged __is_trivial_equality_predicate and
 __is_trivial_plus_operation into __desugars_to

---
 libcxx/include/CMakeLists.txt |  1 -
 libcxx/include/__algorithm/comp.h |  7 ++---
 libcxx/include/__algorithm/equal.h| 22 
 .../cpu_backends/transform_reduce.h   | 23 
 libcxx/include/__functional/operations.h  | 15 +--
 .../include/__functional/ranges_operations.h  |  7 ++---
 .../include/__numeric/pstl_transform_reduce.h |  2 +-
 .../include/__type_traits/operation_traits.h  |  4 +--
 .../include/__type_traits/predicate_traits.h  | 26 ---
 libcxx/include/module.modulemap.in|  1 -
 10 files changed, 40 insertions(+), 68 deletions(-)
 delete mode 100644 libcxx/include/__type_traits/predicate_traits.h

diff --git a/libcxx/include/CMakeLists.txt b/libcxx/include/CMakeLists.txt
index 340353f8ebb41c4..7eb09a06ccd482e 100644
--- a/libcxx/include/CMakeLists.txt
+++ b/libcxx/include/CMakeLists.txt
@@ -813,7 +813,6 @@ set(files
   __type_traits/negation.h
   __type_traits/noexcept_move_assign_container.h
   __type_traits/operation_traits.h
-  __type_traits/predicate_traits.h
   __type_traits/promote.h
   __type_traits/rank.h
   __type_traits/remove_all_extents.h
diff --git a/libcxx/include/__algorithm/comp.h 
b/libcxx/include/__algorithm/comp.h
index 9474536615ffb67..0993c37cce36a6b 100644
--- a/libcxx/include/__algorithm/comp.h
+++ b/libcxx/include/__algorithm/comp.h
@@ -10,8 +10,9 @@
 #define _LIBCPP___ALGORITHM_COMP_H
 
 #include <__config>
+#include <__functional/operations.h>
 #include <__type_traits/integral_constant.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
 #  pragma GCC system_header
@@ -26,8 +27,8 @@ struct __equal_to {
   }
 };
 
-template 
-struct __is_trivial_equality_predicate<__equal_to, _Lhs, _Rhs> : true_type {};
+template <>
+struct __desugars_to<__equal_to, std::equal_to<>> : true_type {};
 
 // The definition is required because __less is part of the ABI, but it's empty
 // because all comparisons should be transparent.
diff --git a/libcxx/include/__algorithm/equal.h 
b/libcxx/include/__algorithm/equal.h
index b69aeff92bb9289..35e82da15e4d058 100644
--- a/libcxx/include/__algorithm/equal.h
+++ b/libcxx/include/__algorithm/equal.h
@@ -15,6 +15,7 @@
 #include <__config>
 #include <__functional/identity.h>
 #include <__functional/invoke.h>
+#include <__functional/operations.h>
 #include <__iterator/distance.h>
 #include <__iterator/iterator_traits.h>
 #include <__string/constexpr_c_functions.h>
@@ -23,7 +24,7 @@
 #include <__type_traits/is_constant_evaluated.h>
 #include <__type_traits/is_equality_comparable.h>
 #include <__type_traits/is_volatile.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 #include <__utility/move.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
@@ -41,13 +42,12 @@ _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI 
_LIBCPP_CONSTEXPR_SINCE_CXX20 boo
   return true;
 }
 
-template <
-class _Tp,
-class _Up,
-class _BinaryPredicate,
-__enable_if_t<__is_trivial_equality_predicate<_BinaryPredicate, _Tp, 
_Up>::value && !is_volatile<_Tp>::value &&
-  !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
-  int> = 0>
+template < class _Tp,
+   class _Up,
+   class _BinaryPredicate,
+   __enable_if_t<__desugars_to<_BinaryPredicate, 
std::equal_to<>>::value && !is_volatile<_Tp>::value &&
+ !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
+ int> = 0>
 _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
 __equal_iter_impl(_Tp* __first1, _Tp* __last1, _Up* __first2, 
_BinaryPredicate&) {
   return std::__constexpr_memcmp_equal(__first1, __first2, 
__element_count(__last1 - __first1));
@@ -94,12 +94,12 @@ template ::value && __is_identity<_Proj1>::value &&
+  __enable_if_t<__desugars_to<_Pred, std::equal_to<>>::value && 
__is_identity<_Proj1>::value &&
 __is_identity<_Proj2>::value && 
!is_volatile<_Tp>::value && !is_volatile<_Up>::value &&
 __libcpp_is_trivially_equality_comparable<_Tp, 
_Up>::value,
 int> = 0>
-_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool __equal_impl(
-_Tp* __first1, _Tp* __last1, _Up* __first2, _Up*, _Pred&, _Proj1&, 
_Proj2&) {
+_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
+

[clang] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/68642

>From f0d93cc6a5cd485c654ab38221691db038bacc7d Mon Sep 17 00:00:00 2001
From: AntonRydahl 
Date: Mon, 9 Oct 2023 15:13:22 -0700
Subject: [PATCH 1/4] Merged __is_trivial_equality_predicate and
 __is_trivial_plus_operation into __desugars_to

---
 libcxx/include/CMakeLists.txt |  1 -
 libcxx/include/__algorithm/comp.h |  7 ++---
 libcxx/include/__algorithm/equal.h| 22 
 .../cpu_backends/transform_reduce.h   | 23 
 libcxx/include/__functional/operations.h  | 15 +--
 .../include/__functional/ranges_operations.h  |  7 ++---
 .../include/__numeric/pstl_transform_reduce.h |  2 +-
 .../include/__type_traits/operation_traits.h  |  4 +--
 .../include/__type_traits/predicate_traits.h  | 26 ---
 libcxx/include/module.modulemap.in|  1 -
 10 files changed, 40 insertions(+), 68 deletions(-)
 delete mode 100644 libcxx/include/__type_traits/predicate_traits.h

diff --git a/libcxx/include/CMakeLists.txt b/libcxx/include/CMakeLists.txt
index 340353f8ebb41c4..7eb09a06ccd482e 100644
--- a/libcxx/include/CMakeLists.txt
+++ b/libcxx/include/CMakeLists.txt
@@ -813,7 +813,6 @@ set(files
   __type_traits/negation.h
   __type_traits/noexcept_move_assign_container.h
   __type_traits/operation_traits.h
-  __type_traits/predicate_traits.h
   __type_traits/promote.h
   __type_traits/rank.h
   __type_traits/remove_all_extents.h
diff --git a/libcxx/include/__algorithm/comp.h 
b/libcxx/include/__algorithm/comp.h
index 9474536615ffb67..0993c37cce36a6b 100644
--- a/libcxx/include/__algorithm/comp.h
+++ b/libcxx/include/__algorithm/comp.h
@@ -10,8 +10,9 @@
 #define _LIBCPP___ALGORITHM_COMP_H
 
 #include <__config>
+#include <__functional/operations.h>
 #include <__type_traits/integral_constant.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
 #  pragma GCC system_header
@@ -26,8 +27,8 @@ struct __equal_to {
   }
 };
 
-template 
-struct __is_trivial_equality_predicate<__equal_to, _Lhs, _Rhs> : true_type {};
+template <>
+struct __desugars_to<__equal_to, std::equal_to<>> : true_type {};
 
 // The definition is required because __less is part of the ABI, but it's empty
 // because all comparisons should be transparent.
diff --git a/libcxx/include/__algorithm/equal.h 
b/libcxx/include/__algorithm/equal.h
index b69aeff92bb9289..35e82da15e4d058 100644
--- a/libcxx/include/__algorithm/equal.h
+++ b/libcxx/include/__algorithm/equal.h
@@ -15,6 +15,7 @@
 #include <__config>
 #include <__functional/identity.h>
 #include <__functional/invoke.h>
+#include <__functional/operations.h>
 #include <__iterator/distance.h>
 #include <__iterator/iterator_traits.h>
 #include <__string/constexpr_c_functions.h>
@@ -23,7 +24,7 @@
 #include <__type_traits/is_constant_evaluated.h>
 #include <__type_traits/is_equality_comparable.h>
 #include <__type_traits/is_volatile.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 #include <__utility/move.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
@@ -41,13 +42,12 @@ _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI 
_LIBCPP_CONSTEXPR_SINCE_CXX20 boo
   return true;
 }
 
-template <
-class _Tp,
-class _Up,
-class _BinaryPredicate,
-__enable_if_t<__is_trivial_equality_predicate<_BinaryPredicate, _Tp, 
_Up>::value && !is_volatile<_Tp>::value &&
-  !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
-  int> = 0>
+template < class _Tp,
+   class _Up,
+   class _BinaryPredicate,
+   __enable_if_t<__desugars_to<_BinaryPredicate, 
std::equal_to<>>::value && !is_volatile<_Tp>::value &&
+ !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
+ int> = 0>
 _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
 __equal_iter_impl(_Tp* __first1, _Tp* __last1, _Up* __first2, 
_BinaryPredicate&) {
   return std::__constexpr_memcmp_equal(__first1, __first2, 
__element_count(__last1 - __first1));
@@ -94,12 +94,12 @@ template ::value && __is_identity<_Proj1>::value &&
+  __enable_if_t<__desugars_to<_Pred, std::equal_to<>>::value && 
__is_identity<_Proj1>::value &&
 __is_identity<_Proj2>::value && 
!is_volatile<_Tp>::value && !is_volatile<_Up>::value &&
 __libcpp_is_trivially_equality_comparable<_Tp, 
_Up>::value,
 int> = 0>
-_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool __equal_impl(
-_Tp* __first1, _Tp* __last1, _Up* __first2, _Up*, _Pred&, _Proj1&, 
_Proj2&) {
+_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
+

[clang-tools-extra] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl updated 
https://github.com/llvm/llvm-project/pull/68642

>From f0d93cc6a5cd485c654ab38221691db038bacc7d Mon Sep 17 00:00:00 2001
From: AntonRydahl 
Date: Mon, 9 Oct 2023 15:13:22 -0700
Subject: [PATCH 1/4] Merged __is_trivial_equality_predicate and
 __is_trivial_plus_operation into __desugars_to

---
 libcxx/include/CMakeLists.txt |  1 -
 libcxx/include/__algorithm/comp.h |  7 ++---
 libcxx/include/__algorithm/equal.h| 22 
 .../cpu_backends/transform_reduce.h   | 23 
 libcxx/include/__functional/operations.h  | 15 +--
 .../include/__functional/ranges_operations.h  |  7 ++---
 .../include/__numeric/pstl_transform_reduce.h |  2 +-
 .../include/__type_traits/operation_traits.h  |  4 +--
 .../include/__type_traits/predicate_traits.h  | 26 ---
 libcxx/include/module.modulemap.in|  1 -
 10 files changed, 40 insertions(+), 68 deletions(-)
 delete mode 100644 libcxx/include/__type_traits/predicate_traits.h

diff --git a/libcxx/include/CMakeLists.txt b/libcxx/include/CMakeLists.txt
index 340353f8ebb41c4..7eb09a06ccd482e 100644
--- a/libcxx/include/CMakeLists.txt
+++ b/libcxx/include/CMakeLists.txt
@@ -813,7 +813,6 @@ set(files
   __type_traits/negation.h
   __type_traits/noexcept_move_assign_container.h
   __type_traits/operation_traits.h
-  __type_traits/predicate_traits.h
   __type_traits/promote.h
   __type_traits/rank.h
   __type_traits/remove_all_extents.h
diff --git a/libcxx/include/__algorithm/comp.h 
b/libcxx/include/__algorithm/comp.h
index 9474536615ffb67..0993c37cce36a6b 100644
--- a/libcxx/include/__algorithm/comp.h
+++ b/libcxx/include/__algorithm/comp.h
@@ -10,8 +10,9 @@
 #define _LIBCPP___ALGORITHM_COMP_H
 
 #include <__config>
+#include <__functional/operations.h>
 #include <__type_traits/integral_constant.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
 #  pragma GCC system_header
@@ -26,8 +27,8 @@ struct __equal_to {
   }
 };
 
-template 
-struct __is_trivial_equality_predicate<__equal_to, _Lhs, _Rhs> : true_type {};
+template <>
+struct __desugars_to<__equal_to, std::equal_to<>> : true_type {};
 
 // The definition is required because __less is part of the ABI, but it's empty
 // because all comparisons should be transparent.
diff --git a/libcxx/include/__algorithm/equal.h 
b/libcxx/include/__algorithm/equal.h
index b69aeff92bb9289..35e82da15e4d058 100644
--- a/libcxx/include/__algorithm/equal.h
+++ b/libcxx/include/__algorithm/equal.h
@@ -15,6 +15,7 @@
 #include <__config>
 #include <__functional/identity.h>
 #include <__functional/invoke.h>
+#include <__functional/operations.h>
 #include <__iterator/distance.h>
 #include <__iterator/iterator_traits.h>
 #include <__string/constexpr_c_functions.h>
@@ -23,7 +24,7 @@
 #include <__type_traits/is_constant_evaluated.h>
 #include <__type_traits/is_equality_comparable.h>
 #include <__type_traits/is_volatile.h>
-#include <__type_traits/predicate_traits.h>
+#include <__type_traits/operation_traits.h>
 #include <__utility/move.h>
 
 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
@@ -41,13 +42,12 @@ _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI 
_LIBCPP_CONSTEXPR_SINCE_CXX20 boo
   return true;
 }
 
-template <
-class _Tp,
-class _Up,
-class _BinaryPredicate,
-__enable_if_t<__is_trivial_equality_predicate<_BinaryPredicate, _Tp, 
_Up>::value && !is_volatile<_Tp>::value &&
-  !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
-  int> = 0>
+template < class _Tp,
+   class _Up,
+   class _BinaryPredicate,
+   __enable_if_t<__desugars_to<_BinaryPredicate, 
std::equal_to<>>::value && !is_volatile<_Tp>::value &&
+ !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
+ int> = 0>
 _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
 __equal_iter_impl(_Tp* __first1, _Tp* __last1, _Up* __first2, 
_BinaryPredicate&) {
   return std::__constexpr_memcmp_equal(__first1, __first2, 
__element_count(__last1 - __first1));
@@ -94,12 +94,12 @@ template ::value && __is_identity<_Proj1>::value &&
+  __enable_if_t<__desugars_to<_Pred, std::equal_to<>>::value && 
__is_identity<_Proj1>::value &&
 __is_identity<_Proj2>::value && 
!is_volatile<_Tp>::value && !is_volatile<_Up>::value &&
 __libcpp_is_trivially_equality_comparable<_Tp, 
_Up>::value,
 int> = 0>
-_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool __equal_impl(
-_Tp* __first1, _Tp* __last1, _Up* __first2, _Up*, _Pred&, _Proj1&, 
_Proj2&) {
+_LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI _LIBCPP_CONSTEXPR_SINCE_CXX20 
bool
+

[clang] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

@@ -41,13 +42,12 @@ _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI 
_LIBCPP_CONSTEXPR_SINCE_CXX20 boo
   return true;
 }
 
-template <
-class _Tp,
-class _Up,
-class _BinaryPredicate,
-__enable_if_t<__is_trivial_equality_predicate<_BinaryPredicate, _Tp, 
_Up>::value && !is_volatile<_Tp>::value &&
-  !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
-  int> = 0>
+template < class _Tp,
+   class _Up,
+   class _BinaryPredicate,
+   __enable_if_t<__desugars_to<_BinaryPredicate, equal_to<_Tp> 
>::value && !is_volatile<_Tp>::value &&

AntonRydahl wrote:

Thanks a lot, Louis! I have tried to verify that the patch produces identical 
assembly for the above example compared to the main branch. Just to make sure 
that we are on the same page, this meant that I also had to remove
```
template 
struct __desugars_to, equal_to > : true_type {};
```
since otherwise all equality operators would be considered transparent.

https://github.com/llvm/llvm-project/pull/68642
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang-tools-extra] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

@@ -41,13 +42,12 @@ _LIBCPP_NODISCARD inline _LIBCPP_HIDE_FROM_ABI 
_LIBCPP_CONSTEXPR_SINCE_CXX20 boo
   return true;
 }
 
-template <
-class _Tp,
-class _Up,
-class _BinaryPredicate,
-__enable_if_t<__is_trivial_equality_predicate<_BinaryPredicate, _Tp, 
_Up>::value && !is_volatile<_Tp>::value &&
-  !is_volatile<_Up>::value && 
__libcpp_is_trivially_equality_comparable<_Tp, _Up>::value,
-  int> = 0>
+template < class _Tp,
+   class _Up,
+   class _BinaryPredicate,
+   __enable_if_t<__desugars_to<_BinaryPredicate, equal_to<_Tp> 
>::value && !is_volatile<_Tp>::value &&

AntonRydahl wrote:

Thanks a lot, Louis! I have tried to verify that the patch produces identical 
assembly for the above example compared to the main branch. Just to make sure 
that we are on the same page, this meant that I also had to remove
```
template 
struct __desugars_to, equal_to > : true_type {};
```
since otherwise all equality operators would be considered transparent.

https://github.com/llvm/llvm-project/pull/68642
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl edited 
https://github.com/llvm/llvm-project/pull/68642
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang-tools-extra] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl edited 
https://github.com/llvm/llvm-project/pull/68642
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl edited 
https://github.com/llvm/llvm-project/pull/68642
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang-tools-extra] [libcxx] Unifying __is_trivial_equality_predicate and __is_trivial_plus_operation into __desugars_to (PR #68642)

[Anton Rydahl via cfe-commits]

https://github.com/AntonRydahl edited 
https://github.com/llvm/llvm-project/pull/68642
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [libc][libm][GPU] Added missing vendor entry points to `libm` for GPUs (PR #66031)

[Anton Rydahl via cfe-commits]

Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,
Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,MaheshRavishankar
 <1663364+maheshravishan...@users.noreply.github.com>,Fangrui Song
 ,Razvan Lupusoru ,Walter Erquinigo
 ,"Yaxun (Sam) Liu" ,Nico Weber
 ,Stella Laurenzo ,Lang Hames
 ,Matthias Braun ,Arthur Eubanks
 ,Nico Weber ,
Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,Vitaly Buka
 ,Vitaly Buka ,Fangrui Song
 ,Fangrui Song ,Fangrui Song 
,Fangrui
 Song ,erichkeane ,Amir Ayupov
 ,Amir Ayupov ,yinying-lisa-li
 <107574043+yinying-lisa...@users.noreply.github.com>,Christopher Di Bella
 ,Konstantin Varlamov ,Alexey Bader
 ,Vitaly Buka 


https://github.com/AntonRydahl labeled 
https://github.com/llvm/llvm-project/pull/66031
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [libc][libm][GPU] Added missing vendor entry points to `libm` for GPUs (PR #66031)

[Anton Rydahl via cfe-commits]

Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,
Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,MaheshRavishankar
 <1663364+maheshravishan...@users.noreply.github.com>,Fangrui Song
 ,Razvan Lupusoru ,Walter Erquinigo
 ,"Yaxun (Sam) Liu" ,Nico Weber
 ,Stella Laurenzo ,Lang Hames
 ,Matthias Braun ,Arthur Eubanks
 ,Nico Weber ,
Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,Vitaly Buka
 ,Vitaly Buka ,Fangrui Song
 ,Fangrui Song ,Fangrui Song 
,Fangrui
 Song ,erichkeane ,Amir Ayupov
 ,Amir Ayupov ,yinying-lisa-li
 <107574043+yinying-lisa...@users.noreply.github.com>,Christopher Di Bella
 ,Konstantin Varlamov ,Alexey Bader
 ,Vitaly Buka 


https://github.com/AntonRydahl review_requested 
https://github.com/llvm/llvm-project/pull/66031
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [libc][libm][GPU] Added missing vendor entry points to `libm` for GPUs (PR #66031)

[Anton Rydahl via cfe-commits]

Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,
Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,MaheshRavishankar
 <1663364+maheshravishan...@users.noreply.github.com>,Fangrui Song
 ,Razvan Lupusoru ,Walter Erquinigo
 ,"Yaxun (Sam) Liu" ,Nico Weber
 ,Stella Laurenzo ,Lang Hames
 ,Matthias Braun ,Arthur Eubanks
 ,Nico Weber ,
Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,Vitaly Buka
 ,Vitaly Buka ,Fangrui Song
 ,Fangrui Song ,Fangrui Song 
,Fangrui
 Song ,erichkeane ,Amir Ayupov
 ,Amir Ayupov ,yinying-lisa-li
 <107574043+yinying-lisa...@users.noreply.github.com>,Christopher Di Bella
 ,Konstantin Varlamov ,Alexey Bader
 ,Vitaly Buka 


https://github.com/AntonRydahl review_requested 
https://github.com/llvm/llvm-project/pull/66031
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [libc][libm][GPU] Added missing vendor entry points to `libm` for GPUs (PR #66031)

[Anton Rydahl via cfe-commits]

Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,
Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,MaheshRavishankar
 <1663364+maheshravishan...@users.noreply.github.com>,Fangrui Song
 ,Razvan Lupusoru ,Walter Erquinigo
 ,"Yaxun (Sam) Liu" ,Nico Weber
 ,Stella Laurenzo ,Lang Hames
 ,Matthias Braun ,Arthur Eubanks
 ,Nico Weber ,
Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,Vitaly Buka
 ,Vitaly Buka ,Fangrui Song
 ,Fangrui Song ,Fangrui Song 
,Fangrui
 Song ,erichkeane ,Amir Ayupov
 ,Amir Ayupov ,yinying-lisa-li
 <107574043+yinying-lisa...@users.noreply.github.com>,Christopher Di Bella
 ,Konstantin Varlamov ,Alexey Bader
 ,Vitaly Buka 


https://github.com/AntonRydahl review_requested 
https://github.com/llvm/llvm-project/pull/66031
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [libc][libm][GPU] Added missing vendor entry points to `libm` for GPUs (PR #66031)

[Anton Rydahl via cfe-commits]

Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,
Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,MaheshRavishankar
 <1663364+maheshravishan...@users.noreply.github.com>,Fangrui Song
 ,Razvan Lupusoru ,Walter Erquinigo
 ,"Yaxun (Sam) Liu" ,Nico Weber
 ,Stella Laurenzo ,Lang Hames
 ,Matthias Braun ,Arthur Eubanks
 ,Nico Weber ,
Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,Vitaly Buka
 ,Vitaly Buka ,Fangrui Song
 ,Fangrui Song ,Fangrui Song 
,Fangrui
 Song ,erichkeane ,Amir Ayupov
 ,Amir Ayupov ,yinying-lisa-li
 <107574043+yinying-lisa...@users.noreply.github.com>,Christopher Di Bella
 ,Konstantin Varlamov ,Alexey Bader
 ,Vitaly Buka 


https://github.com/AntonRydahl review_requested 
https://github.com/llvm/llvm-project/pull/66031
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [libc][libm][GPU] Added missing vendor entry points to `libm` for GPUs (PR #66031)

[Anton Rydahl via cfe-commits]

Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,
Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,MaheshRavishankar
 <1663364+maheshravishan...@users.noreply.github.com>,Fangrui Song
 ,Razvan Lupusoru ,Walter Erquinigo
 ,"Yaxun (Sam) Liu" ,Nico Weber
 ,Stella Laurenzo ,Lang Hames
 ,Matthias Braun ,Arthur Eubanks
 ,Nico Weber ,
Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,Vitaly Buka
 ,Vitaly Buka ,Fangrui Song
 ,Fangrui Song ,Fangrui Song 
,Fangrui
 Song ,erichkeane ,Amir Ayupov
 ,Amir Ayupov ,yinying-lisa-li
 <107574043+yinying-lisa...@users.noreply.github.com>,Christopher Di Bella
 ,Konstantin Varlamov ,Alexey Bader
 ,Vitaly Buka 


https://github.com/AntonRydahl closed 
https://github.com/llvm/llvm-project/pull/66031
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits