NAME¶

std::atomic::fetch_add - std::atomic::fetch_add

Synopsis¶

member only of atomic<Integral > specializations
and atomic<Floating > specializations
(since C++20)
T fetch_add( T arg, std::memory_order order = (1) (since C++11)
std::memory_order_seq_cst ) noexcept;
T fetch_add( T arg, std::memory_order order =
std::memory_order_seq_cst ) volatile (2) (since C++11)
noexcept;
member only of atomic<T*> partial specialization
T* fetch_add( std::ptrdiff_t arg,

std::memory_order order = (3) (since C++11)

std::memory_order_seq_cst ) noexcept;
T* fetch_add( std::ptrdiff_t arg,

std::memory_order order = (4) (since C++11)

std::memory_order_seq_cst ) volatile noexcept;

Atomically replaces the current value with the result of arithmetic addition of the
value and arg. That is, it performs atomic post-increment. The operation is a
read-modify-write operation. Memory is affected according to the value of order.

1,2) For signed integral types, arithmetic is defined to use two’s complement
representation. There are no undefined results.

For floating-point types, the floating-point environment in effect may
be different from the calling thread's floating-point environment. The
operation need not conform to the corresponding std::numeric_limits (since C++20)
traits but is encouraged to do so. If the result is not a
representable value for its type, the result is unspecified but the
operation otherwise has no undefined behavior.

3,4) The result may be an undefined address, but the operation otherwise has no
undefined behavior.
If T is not a complete object type, the program is ill-formed.

It is deprecated if std::atomic<T>::is_always_lock_free is false and (since C++20)
overload (2) or (4) participates in overload resolution.

Parameters¶

arg - the other argument of arithmetic addition
order - memory order constraints to enforce

Return value¶

The value immediately preceding the effects of this function in the modification
order of *this.

Example¶

// Run this code

#include <array>
#include <atomic>
#include <iostream>
#include <thread>

std::atomic<long long> data{10};
std::array<long long, 5> return_values{};

void do_work(int thread_num)
{
long long val = data.fetch_add(1, std::memory_order_relaxed);
return_values[thread_num] = val;
}

int main()
{
{
std::jthread th0{do_work, 0};
std::jthread th1{do_work, 1};
std::jthread th2{do_work, 2};
std::jthread th3{do_work, 3};
std::jthread th4{do_work, 4};
}

std::cout << "Result : " << data << '\n';

for (long long val : return_values)
std::cout << "Seen return value : " << val << std::endl;
}

Possible output:¶

Result : 15
Seen return value : 11
Seen return value : 10
Seen return value : 14
Seen return value : 12
Seen return value : 13

Defect reports

The following behavior-changing defect reports were applied retroactively to
previously published C++ standards.

DR Applied to Behavior as published Correct behavior
P0558R1 C++11 arithmetic permitted on pointers to (possibly made ill-formed
cv-qualified) void or function

See also¶

atomic_fetch_add adds a non-atomic value to an atomic object and obtains
atomic_fetch_add_explicit the previous value of the atomic
(C++11) (function template)
(C++11)
operator++
operator++(int) increments or decrements the atomic value by one
operator-- (public member function)
operator--(int)