cuda_toolkit/include/support/atomic/atomic_msvc.h

// -*- C++ -*-
//===----------------------------------------------------------------------===//
//
// Part of libcu++, the C++ Standard Library for your entire system,
// 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
// SPDX-FileCopyrightText: Copyright (c) 2023 NVIDIA CORPORATION & AFFILIATES.
//
//===----------------------------------------------------------------------===//

#ifndef _MSC_VER
#error "This file is only for CL.EXE's benefit"
#endif

#define _LIBCUDACXX_COMPILER_BARRIER() _ReadWriteBarrier()

#if defined(_M_ARM) || defined(_M_ARM64)
    #define _LIBCUDACXX_MEMORY_BARRIER()             __dmb(0xB) // inner shared data memory barrier
    #define _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER() _LIBCUDACXX_MEMORY_BARRIER()
#elif defined(_M_IX86) || defined(_M_X64)
    #define _LIBCUDACXX_MEMORY_BARRIER()             __faststorefence()
    // x86/x64 hardware only emits memory barriers inside _Interlocked intrinsics
    #define _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER() _LIBCUDACXX_COMPILER_BARRIER()
#else // ^^^ x86/x64 / unsupported hardware vvv
    #error Unsupported hardware
#endif // hardware

// MSVC Does not have compiler intrinsics for lock-free checking
#ifndef _LIBCUDACXX_ATOMIC_IS_LOCK_FREE
#define _LIBCUDACXX_ATOMIC_IS_LOCK_FREE(__x) (__x <= 8)
#endif

inline int __stronger_order_msvc(int __a, int __b) {
    int const __max = __a > __b ? __a : __b;
    if(__max != __ATOMIC_RELEASE)
        return __max;
    static int const __xform[] = {
        __ATOMIC_RELEASE,
        __ATOMIC_ACQ_REL,
        __ATOMIC_ACQ_REL,
        __ATOMIC_RELEASE };
    return __xform[__a < __b ? __a : __b];
}

static inline void __atomic_signal_fence(int __memorder) {
    if (__memorder != __ATOMIC_RELAXED)
        _LIBCUDACXX_COMPILER_BARRIER();
}

static inline void __atomic_thread_fence(int __memorder) {
    if (__memorder != __ATOMIC_RELAXED)
        _LIBCUDACXX_MEMORY_BARRIER();
}

template <typename _Type, size_t _Size>
using _enable_if_sized_as = typename enable_if<sizeof(_Type) == _Size, int>::type;

template<class _Type, _enable_if_sized_as<_Type,1> = 0>
void __atomic_load_relaxed(const volatile _Type *__ptr, _Type *__ret) {
#ifdef _LIBCUDACXX_MSVC_HAS_NO_ISO_INTRIN
    __int8 __tmp = *(const volatile __int8 *)__ptr;
#else
    __int8 __tmp = __iso_volatile_load8((const volatile __int8 *)__ptr);
#endif
    *__ret = reinterpret_cast<_Type&>(__tmp);
}
template<class _Type, _enable_if_sized_as<_Type,2> = 0>
void __atomic_load_relaxed(const volatile _Type *__ptr, _Type *__ret) {
#ifdef _LIBCUDACXX_MSVC_HAS_NO_ISO_INTRIN
    __int16 __tmp = *(const volatile __int16 *)__ptr;
#else
    __int16 __tmp = __iso_volatile_load16((const volatile __int16 *)__ptr);
#endif
    *__ret = reinterpret_cast<_Type&>(__tmp);
}
template<class _Type, _enable_if_sized_as<_Type,4> = 0>
void __atomic_load_relaxed(const volatile _Type *__ptr, _Type *__ret) {
#ifdef _LIBCUDACXX_MSVC_HAS_NO_ISO_INTRIN
    __int32 __tmp = *(const volatile __int32 *)__ptr;
#else
    __int32 __tmp = __iso_volatile_load32((const volatile __int32 *)__ptr);
#endif
    *__ret = reinterpret_cast<_Type&>(__tmp);
}
template<class _Type, _enable_if_sized_as<_Type,8> = 0>
void __atomic_load_relaxed(const volatile _Type *__ptr, _Type *__ret) {
#ifdef _LIBCUDACXX_MSVC_HAS_NO_ISO_INTRIN
    __int64 __tmp = *(const volatile __int64 *)__ptr;
#else
    __int64 __tmp = __iso_volatile_load64((const volatile __int64 *)__ptr);
#endif
    *__ret = reinterpret_cast<_Type&>(__tmp);
}

template<class _Type>
void __atomic_load(const volatile _Type *__ptr, _Type *__ret, int __memorder) {
    switch (__memorder) {
    case __ATOMIC_SEQ_CST: _LIBCUDACXX_MEMORY_BARRIER(); _LIBCUDACXX_FALLTHROUGH();
    case __ATOMIC_CONSUME:
    case __ATOMIC_ACQUIRE: __atomic_load_relaxed(__ptr, __ret); _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); break;
    case __ATOMIC_RELAXED: __atomic_load_relaxed(__ptr, __ret); break;
    default: assert(0);
    }
}

template<class _Type, _enable_if_sized_as<_Type,1> = 0>
void __atomic_store_relaxed(volatile _Type *__ptr, _Type *__val) {
    auto __t = reinterpret_cast<__int8 *>(__val);
    auto __d = reinterpret_cast<volatile __int8 *>(__ptr);
#ifdef _LIBCUDACXX_MSVC_HAS_NO_ISO_INTRIN
    (void)_InterlockedExchange8(__d, *__t);
#else
    __iso_volatile_store8(__d, *__t);
#endif
}
template<class _Type, _enable_if_sized_as<_Type,2> = 0>
void __atomic_store_relaxed(volatile _Type *__ptr, _Type *__val) {
    auto __t = reinterpret_cast<__int16 *>(__val);
    auto __d = reinterpret_cast<volatile __int16 *>(__ptr);
#ifdef _LIBCUDACXX_MSVC_HAS_NO_ISO_INTRIN
    (void)_InterlockedExchange16(__d, *__t);
#else
    __iso_volatile_store16(__d, *__t);
#endif
}
template<class _Type, _enable_if_sized_as<_Type,4> = 0>
void __atomic_store_relaxed(volatile _Type *__ptr, _Type *__val) {
    auto __t = reinterpret_cast<__int32 *>(__val);
    auto __d = reinterpret_cast<volatile __int32 *>(__ptr);
#ifdef _LIBCUDACXX_MSVC_HAS_NO_ISO_INTRIN
    // int cannot be converted to long?...
    (void)_InterlockedExchange(reinterpret_cast<volatile long *>(__d), *__t);
#else
    __iso_volatile_store32(__d, *__t);
#endif
}
template<class _Type, _enable_if_sized_as<_Type,8> = 0>
void __atomic_store_relaxed(volatile _Type *__ptr, _Type *__val) {
    auto __t = reinterpret_cast<__int64 *>(__val);
    auto __d = reinterpret_cast<volatile __int64 *>(__ptr);
#ifdef _LIBCUDACXX_MSVC_HAS_NO_ISO_INTRIN
    (void)_InterlockedExchange64(__d, *__t);
#else
    __iso_volatile_store64(__d, *__t);
#endif
}

template<class _Type>
void __atomic_store(volatile _Type *__ptr, _Type *__val, int __memorder) {
    switch (__memorder) {
    case __ATOMIC_RELEASE: _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); __atomic_store_relaxed(__ptr, __val); break;
    case __ATOMIC_SEQ_CST: _LIBCUDACXX_MEMORY_BARRIER(); _LIBCUDACXX_FALLTHROUGH();
    case __ATOMIC_RELAXED: __atomic_store_relaxed(__ptr, __val); break;
    default: assert(0);
    }
}

template<class _Type, _enable_if_sized_as<_Type,1> = 0>
bool __atomic_compare_exchange_relaxed(const volatile _Type *__ptr, _Type *__expected, const _Type *__desired) {
    auto __tmp_desired = reinterpret_cast<const char&>(*__desired);
    auto __tmp_expected = reinterpret_cast<char&>(*__expected);
    auto const __old = _InterlockedCompareExchange8((volatile char *)__ptr, __tmp_desired, __tmp_expected);
    if(__old == __tmp_expected)
        return true;
    *__expected = reinterpret_cast<const _Type&>(__old);
    return false;
}
template<class _Type, _enable_if_sized_as<_Type,2> = 0>
bool __atomic_compare_exchange_relaxed(const volatile _Type *__ptr, _Type *__expected, const _Type *__desired) {
    auto __tmp_desired = reinterpret_cast<const short&>(*__desired);
    auto __tmp_expected = reinterpret_cast<short&>(*__expected);
    auto const __old = _InterlockedCompareExchange16((volatile short *)__ptr, __tmp_desired, __tmp_expected);
    if(__old == __tmp_expected)
        return true;
    *__expected = reinterpret_cast<const _Type&>(__old);
    return false;
}
template<class _Type, _enable_if_sized_as<_Type,4> = 0>
bool __atomic_compare_exchange_relaxed(const volatile _Type *__ptr, _Type *__expected, const _Type *__desired) {
    auto __tmp_desired = reinterpret_cast<const long&>(*__desired);
    auto __tmp_expected = reinterpret_cast<long&>(*__expected);
    auto const __old = _InterlockedCompareExchange((volatile long *)__ptr, __tmp_desired, __tmp_expected);
    if(__old == __tmp_expected)
        return true;
    *__expected = reinterpret_cast<const _Type&>(__old);
    return false;
}
template<class _Type, _enable_if_sized_as<_Type,8> = 0>
bool __atomic_compare_exchange_relaxed(const volatile _Type *__ptr, _Type *__expected, const _Type *__desired) {
    auto __tmp_desired = reinterpret_cast<const __int64&>(*__desired);
    auto __tmp_expected = reinterpret_cast<__int64&>(*__expected);
    auto const __old = _InterlockedCompareExchange64((volatile __int64 *)__ptr, __tmp_desired, __tmp_expected);
    if(__old == __tmp_expected)
        return true;
    *__expected = reinterpret_cast<const _Type&>(__old);
    return false;
}
template<class _Type>
bool __atomic_compare_exchange(_Type volatile *__ptr, _Type *__expected, const _Type *__desired, bool, int __success_memorder, int __failure_memorder) {
    bool success = false;
    switch (__stronger_order_msvc(__success_memorder, __failure_memorder)) {
    case __ATOMIC_RELEASE: _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); success = __atomic_compare_exchange_relaxed(__ptr, __expected, __desired); break;
    case __ATOMIC_ACQ_REL: _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); _LIBCUDACXX_FALLTHROUGH();
    case __ATOMIC_CONSUME:
    case __ATOMIC_ACQUIRE: success = __atomic_compare_exchange_relaxed(__ptr, __expected, __desired); _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); break;
    case __ATOMIC_SEQ_CST: _LIBCUDACXX_MEMORY_BARRIER(); success = __atomic_compare_exchange_relaxed(__ptr, __expected, __desired); _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); break;
    case __ATOMIC_RELAXED: success = __atomic_compare_exchange_relaxed(__ptr, __expected, __desired); break;
    default: assert(0);
    }
    return success;
}

template<class _Type, _enable_if_sized_as<_Type,1> = 0>
void __atomic_exchange_relaxed(const volatile _Type *__ptr, const _Type *__val, _Type *__ret) {
    auto const __old = _InterlockedExchange8((volatile char *)__ptr, reinterpret_cast<char const&>(*__val));
    *__ret = reinterpret_cast<_Type const&>(__old);
}
template<class _Type, _enable_if_sized_as<_Type,2> = 0>
void __atomic_exchange_relaxed(const volatile _Type *__ptr, const _Type *__val, _Type *__ret) {
    auto const __old = _InterlockedExchange16((volatile short *)__ptr, reinterpret_cast<short const&>(*__val));
    *__ret = reinterpret_cast<_Type const&>(__old);
}
template<class _Type, _enable_if_sized_as<_Type,4> = 0>
void __atomic_exchange_relaxed(const volatile _Type *__ptr, const _Type *__val, _Type *__ret) {
    auto const __old = _InterlockedExchange((volatile long *)__ptr, reinterpret_cast<long const&>(*__val));
    *__ret = reinterpret_cast<_Type const&>(__old);
}
template<class _Type, _enable_if_sized_as<_Type,8> = 0>
void __atomic_exchange_relaxed(const volatile _Type *__ptr, const _Type *__val, _Type *__ret) {
    auto const __old = _InterlockedExchange64((volatile __int64 *)__ptr, reinterpret_cast<__int64 const&>(*__val));
    *__ret = reinterpret_cast<_Type const&>(__old);
}
template<class _Type>
void __atomic_exchange(_Type volatile *__ptr, const _Type *__val, _Type *__ret, int __memorder) {
    switch (__memorder) {
    case __ATOMIC_RELEASE: _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); __atomic_exchange_relaxed(__ptr, __val, __ret);break;
    case __ATOMIC_ACQ_REL: _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); _LIBCUDACXX_FALLTHROUGH();
    case __ATOMIC_CONSUME:
    case __ATOMIC_ACQUIRE: __atomic_exchange_relaxed(__ptr, __val, __ret); _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); break;
    case __ATOMIC_SEQ_CST: _LIBCUDACXX_MEMORY_BARRIER(); __atomic_exchange_relaxed(__ptr, __val, __ret); _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); break;
    case __ATOMIC_RELAXED: __atomic_exchange_relaxed(__ptr, __val, __ret); break;
    default: assert(0);
    }
}

template<class _Type, class _Delta, _enable_if_sized_as<_Type,1> = 0>
void __atomic_fetch_add_relaxed(const volatile _Type *__ptr, const _Delta *__val, _Type *__ret) {
    auto const __old = _InterlockedExchangeAdd8((volatile char *)__ptr, reinterpret_cast<char const&>(*__val));
    *__ret = reinterpret_cast<_Type const&>(__old);
}
template<class _Type, class _Delta, _enable_if_sized_as<_Type,2> = 0>
void __atomic_fetch_add_relaxed(const volatile _Type *__ptr, const _Delta *__val, _Type *__ret) {
    auto const __old = _InterlockedExchangeAdd16((volatile short *)__ptr, reinterpret_cast<short const&>(*__val));
    *__ret = reinterpret_cast<_Type const&>(__old);
}
template<class _Type, class _Delta, _enable_if_sized_as<_Type,4> = 0>
void __atomic_fetch_add_relaxed(const volatile _Type *__ptr, const _Delta *__val, _Type *__ret) {
    auto const __old = _InterlockedExchangeAdd((volatile long *)__ptr, reinterpret_cast<long const&>(*__val));
    *__ret = reinterpret_cast<_Type const&>(__old);
}
template<class _Type, class _Delta, _enable_if_sized_as<_Type,8> = 0>
void __atomic_fetch_add_relaxed(const volatile _Type *__ptr, const _Delta *__val, _Type *__ret) {
    auto const __old = _InterlockedExchangeAdd64((volatile __int64 *)__ptr, reinterpret_cast<__int64 const&>(*__val));
    *__ret = reinterpret_cast<_Type const&>(__old);
}
template<class _Type, class _Delta>
_Type __atomic_fetch_add(_Type volatile *__ptr, _Delta __val, int __memorder) {
    alignas(_Type) unsigned char __buf[sizeof(_Type)] = {};
    auto* __dest = reinterpret_cast<_Type*>(__buf);

    switch (__memorder) {
    case __ATOMIC_RELEASE: _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); __atomic_fetch_add_relaxed(__ptr, &__val, __dest);break;
    case __ATOMIC_ACQ_REL: _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); _LIBCUDACXX_FALLTHROUGH();
    case __ATOMIC_CONSUME:
    case __ATOMIC_ACQUIRE: __atomic_fetch_add_relaxed(__ptr, &__val, __dest); _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); break;
    case __ATOMIC_SEQ_CST: _LIBCUDACXX_MEMORY_BARRIER(); __atomic_fetch_add_relaxed(__ptr, &__val, __dest); _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); break;
    case __ATOMIC_RELAXED: __atomic_fetch_add_relaxed(__ptr, &__val, __dest); break;
    default: assert(0);
    }
    return *__dest;
}
template<class _Type, class _Delta>
_Type __atomic_fetch_sub(_Type volatile *__ptr, _Delta __val, int __memorder) {
    return __atomic_fetch_add(__ptr, 0-__val, __memorder);
}


template<class _Type, class _Delta, _enable_if_sized_as<_Type,1> = 0>
void __atomic_fetch_and_relaxed(const volatile _Type *__ptr, const _Delta *__val, _Type *__ret) {
    auto const __old = _InterlockedAnd8((volatile char *)__ptr, reinterpret_cast<char const&>(*__val));
    *__ret = reinterpret_cast<_Type const&>(__old);
}
template<class _Type, class _Delta, _enable_if_sized_as<_Type,2> = 0>
void __atomic_fetch_and_relaxed(const volatile _Type *__ptr, const _Delta *__val, _Type *__ret) {
    auto const __old = _InterlockedAnd16((volatile short *)__ptr, reinterpret_cast<short const&>(*__val));
    *__ret = reinterpret_cast<_Type const&>(__old);
}
template<class _Type, class _Delta, _enable_if_sized_as<_Type,4> = 0>
void __atomic_fetch_and_relaxed(const volatile _Type *__ptr, const _Delta *__val, _Type *__ret) {
    auto const __old = _InterlockedAnd((volatile long *)__ptr, reinterpret_cast<long const&>(*__val));
    *__ret = reinterpret_cast<_Type const&>(__old);
}
template<class _Type, class _Delta, _enable_if_sized_as<_Type,8> = 0>
void __atomic_fetch_and_relaxed(const volatile _Type *__ptr, const _Delta *__val, _Type *__ret) {
    auto const __old = _InterlockedAnd64((volatile __int64 *)__ptr, reinterpret_cast<__int64 const&>(*__val));
    *__ret = reinterpret_cast<_Type const&>(__old);
}
template<class _Type, class _Delta>
_Type __atomic_fetch_and(_Type volatile *__ptr, _Delta __val, int __memorder) {
    alignas(_Type) unsigned char __buf[sizeof(_Type)] = {};
    auto* __dest = reinterpret_cast<_Type*>(__buf);

    switch (__memorder) {
    case __ATOMIC_RELEASE: _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); __atomic_fetch_and_relaxed(__ptr, &__val, __dest);break;
    case __ATOMIC_ACQ_REL: _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); _LIBCUDACXX_FALLTHROUGH();
    case __ATOMIC_CONSUME:
    case __ATOMIC_ACQUIRE: __atomic_fetch_and_relaxed(__ptr, &__val, __dest); _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); break;
    case __ATOMIC_SEQ_CST: _LIBCUDACXX_MEMORY_BARRIER(); __atomic_fetch_and_relaxed(__ptr, &__val, __dest); _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); break;
    case __ATOMIC_RELAXED: __atomic_fetch_and_relaxed(__ptr, &__val, __dest); break;
    default: assert(0);
    }
    return *__dest;
}

template<class _Type, class _Delta, _enable_if_sized_as<_Type,1> = 0>
void __atomic_fetch_xor_relaxed(const volatile _Type *__ptr, const _Delta *__val, _Type *__ret) {
    auto const __old = _InterlockedXor8((volatile char *)__ptr, reinterpret_cast<char const&>(*__val));
    *__ret = reinterpret_cast<_Type const&>(__old);
}
template<class _Type, class _Delta, _enable_if_sized_as<_Type,2> = 0>
void __atomic_fetch_xor_relaxed(const volatile _Type *__ptr, const _Delta *__val, _Type *__ret) {
    auto const __old = _InterlockedXor16((volatile short *)__ptr, reinterpret_cast<short const&>(*__val));
    *__ret = reinterpret_cast<_Type const&>(__old);
}
template<class _Type, class _Delta, _enable_if_sized_as<_Type,4> = 0>
void __atomic_fetch_xor_relaxed(const volatile _Type *__ptr, const _Delta *__val, _Type *__ret) {
    auto const __old = _InterlockedXor((volatile long *)__ptr, reinterpret_cast<long const&>(*__val));
    *__ret = reinterpret_cast<_Type const&>(__old);
}
template<class _Type, class _Delta, _enable_if_sized_as<_Type,8> = 0>
void __atomic_fetch_xor_relaxed(const volatile _Type *__ptr, const _Delta *__val, _Type *__ret) {
    auto const __old = _InterlockedXor64((volatile __int64 *)__ptr, reinterpret_cast<__int64 const&>(*__val));
    *__ret = reinterpret_cast<_Type const&>(__old);
}
template<class _Type, class _Delta>
_Type __atomic_fetch_xor(_Type volatile *__ptr, _Delta __val, int __memorder) {
    alignas(_Type) unsigned char __buf[sizeof(_Type)] = {};
    auto* __dest = reinterpret_cast<_Type*>(__buf);

    switch (__memorder) {
    case __ATOMIC_RELEASE: _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); __atomic_fetch_xor_relaxed(__ptr, &__val, __dest);break;
    case __ATOMIC_ACQ_REL: _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); _LIBCUDACXX_FALLTHROUGH();
    case __ATOMIC_CONSUME:
    case __ATOMIC_ACQUIRE: __atomic_fetch_xor_relaxed(__ptr, &__val, __dest); _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); break;
    case __ATOMIC_SEQ_CST: _LIBCUDACXX_MEMORY_BARRIER(); __atomic_fetch_xor_relaxed(__ptr, &__val, __dest); _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); break;
    case __ATOMIC_RELAXED: __atomic_fetch_xor_relaxed(__ptr, &__val, __dest); break;
    default: assert(0);
    }
    return *__dest;
}

template<class _Type, class _Delta, _enable_if_sized_as<_Type,1> = 0>
void __atomic_fetch_or_relaxed(const volatile _Type *__ptr, const _Delta *__val, _Type *__ret) {
    auto const __old = _InterlockedOr8((volatile char *)__ptr, reinterpret_cast<char const&>(*__val));
    *__ret = reinterpret_cast<_Type const&>(__old);
}
template<class _Type, class _Delta, _enable_if_sized_as<_Type,2> = 0>
void __atomic_fetch_or_relaxed(const volatile _Type *__ptr, const _Delta *__val, _Type *__ret) {
    auto const __old = _InterlockedOr16((volatile short *)__ptr, reinterpret_cast<short const&>(*__val));
    *__ret = reinterpret_cast<_Type const&>(__old);
}
template<class _Type, class _Delta, _enable_if_sized_as<_Type,4> = 0>
void __atomic_fetch_or_relaxed(const volatile _Type *__ptr, const _Delta *__val, _Type *__ret) {
    auto const __old = _InterlockedOr((volatile long *)__ptr, reinterpret_cast<long const&>(*__val));
    *__ret = reinterpret_cast<_Type const&>(__old);
}
template<class _Type, class _Delta, _enable_if_sized_as<_Type,8> = 0>
void __atomic_fetch_or_relaxed(const volatile _Type *__ptr, const _Delta *__val, _Type *__ret) {
    auto const __old = _InterlockedOr64((volatile __int64 *)__ptr, reinterpret_cast<__int64 const&>(*__val));
    *__ret = reinterpret_cast<_Type const&>(__old);
}
template<class _Type, class _Delta>
_Type __atomic_fetch_or(_Type volatile *__ptr, _Delta __val, int __memorder) {
    alignas(_Type) unsigned char __buf[sizeof(_Type)] = {};
    auto* __dest = reinterpret_cast<_Type*>(__buf);

    switch (__memorder) {
    case __ATOMIC_RELEASE: _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); __atomic_fetch_or_relaxed(__ptr, &__val, __dest);break;
    case __ATOMIC_ACQ_REL: _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); _LIBCUDACXX_FALLTHROUGH();
    case __ATOMIC_CONSUME:
    case __ATOMIC_ACQUIRE: __atomic_fetch_or_relaxed(__ptr, &__val, __dest); _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); break;
    case __ATOMIC_SEQ_CST: _LIBCUDACXX_MEMORY_BARRIER(); __atomic_fetch_or_relaxed(__ptr, &__val, __dest); _LIBCUDACXX_COMPILER_OR_MEMORY_BARRIER(); break;
    case __ATOMIC_RELAXED: __atomic_fetch_or_relaxed(__ptr, &__val, __dest); break;
    default: assert(0);
    }
    return *__dest;
}

template<class _Type>
_Type __atomic_load_n(const _Type volatile *__ptr, int __memorder) {
    alignas(_Type) unsigned char __buf[sizeof(_Type)] = {};
    auto* __dest = reinterpret_cast<_Type*>(__buf);

    __atomic_load(__ptr, __dest, __memorder);
    return *__dest;
}

template<class _Type>
void __atomic_store_n(_Type volatile *__ptr, _Type __val, int __memorder) {
    __atomic_store(__ptr, &__val, __memorder);
}

template<class _Type>
bool __atomic_compare_exchange_n(_Type volatile *__ptr, _Type *__expected, _Type __desired, bool __weak, int __success_memorder, int __failure_memorder) {
    return __atomic_compare_exchange(__ptr, __expected, &__desired, __weak, __success_memorder, __failure_memorder);
}

template<class _Type>
_Type __atomic_exchange_n(_Type volatile *__ptr, _Type __val, int __memorder) {
    alignas(_Type) unsigned char __buf[sizeof(_Type)] = {};
    auto* __dest = reinterpret_cast<_Type*>(__buf);

    __atomic_exchange(__ptr, &__val, __dest, __memorder);
    return *__dest;
}

template<class _Type, class _Delta>
_Type __atomic_fetch_max(_Type volatile *__ptr, _Delta __val, int __memorder) {
    _Type __expected = __atomic_load_n(__ptr, __ATOMIC_RELAXED);
    _Type __desired = __expected < __val ? __expected : __val;
    while(__desired == __val &&
          !__atomic_compare_exchange_n(__ptr, &__expected, __desired, __memorder, __memorder)) {
        __desired = __expected > __val ? __expected : __val;
    }
    return __expected;
}

template<class _Type, class _Delta>
_Type __atomic_fetch_min(_Type volatile *__ptr, _Delta __val, int __memorder) {
    _Type __expected = __atomic_load_n(__ptr, __ATOMIC_RELAXED);
    _Type __desired = __expected < __val ? __expected : __val;
    while(__desired != __val &&
          !__atomic_compare_exchange_n(__ptr, &__expected, __desired, __memorder, __memorder)) {
        __desired = __expected < __val ? __expected : __val;
    }
    return __expected;
}

#include "atomic_base.h"