std:: atomic_thread_fence

Establishes memory synchronization ordering of non-atomic and relaxed atomic accesses, as instructed by order , without an associated atomic operation. Note however, that at least one atomic operation is required to set up the synchronization, as described below.

Fence-atomic synchronization

A release fence F in thread A synchronizes-with atomic acquire operation Y in thread B , if

• there exists an atomic store X (with any memory order),
• Y reads the value written by X (or the value would be written by release sequence headed by X if X were a release operation),
• F is sequenced-before X in thread A .

In this case, all non-atomic and relaxed atomic stores that are sequenced-before F in thread A will happen-before all non-atomic and relaxed atomic loads from the same locations made in thread B after Y .

Atomic-fence synchronization

An atomic release operation X in thread A synchronizes-with an acquire fence F in thread B , if

• there exists an atomic read Y (with any memory order),
• Y reads the value written by X (or by the release sequence headed by X ),
• Y is sequenced-before F in thread B .

In this case, all non-atomic and relaxed atomic stores that are sequenced-before X in thread A will happen-before all non-atomic and relaxed atomic loads from the same locations made in thread B after F .

Fence-fence synchronization

A release fence FA in thread A synchronizes-with an acquire fence FB in thread B , if

• there exists an atomic object M ,
• there exists an atomic write X (with any memory order) that modifies M in thread A ,
• FA is sequenced-before X in thread A ,
• there exists an atomic read Y (with any memory order) in thread B ,
• Y reads the value written by X (or the value would be written by release sequence headed by X if X were a release operation),
• Y is sequenced-before FB in thread B .

In this case, all non-atomic and relaxed atomic stores that are sequenced-before FA in thread A will happen-before all non-atomic and relaxed atomic loads from the same locations made in thread B after FB .

Depending on the value of the order parameter, the effects of this call are:

• When order == std:: memory_order_relaxed , there are no effects.
• When order == std:: memory_order_acquire or order == std:: memory_order_consume , is an acquire fence.
• When order == std:: memory_order_release , is a release fence.
• When order == std:: memory_order_acq_rel , is both a release fence and an acquire fence.
• When order == std:: memory_order_seq_cst , is a sequentially-consistent ordering acquire fence and release fence.

Parameters

Notes

On x86 (including x86-64), atomic_thread_fence functions issue no CPU instructions and only affect compile-time code motion, except for std :: atomic_thread_fence ( std:: memory_order_seq_cst ) .

atomic_thread_fence imposes stronger synchronization constraints than an atomic store operation with the same std::memory_order . While an atomic store-release operation prevents all preceding reads and writes from moving past the store-release, an atomic_thread_fence with std:: memory_order_release ordering prevents all preceding reads and writes from moving past all subsequent stores.

Fence-fence synchronization can be used to add synchronization to a sequence of several relaxed atomic operations, for example:

// Global
std::string computation(int);
void print(std::string);
std::atomic<int> arr[3] = {-1, -1, -1};
std::string data[1000]; //non-atomic data
// Thread A, compute 3 values.
void ThreadA(int v0, int v1, int v2)
{
//  assert(0 <= v0, v1, v2 < 1000);
    data[v0] = computation(v0);
    data[v1] = computation(v1);
    data[v2] = computation(v2);
    std::atomic_thread_fence(std::memory_order_release);
    std::atomic_store_explicit(&arr[0], v0, std::memory_order_relaxed);
    std::atomic_store_explicit(&arr[1], v1, std::memory_order_relaxed);
    std::atomic_store_explicit(&arr[2], v2, std::memory_order_relaxed);
}
// Thread B, prints between 0 and 3 values already computed.
void ThreadB()
{
    int v0 = std::atomic_load_explicit(&arr[0], std::memory_order_relaxed);
    int v1 = std::atomic_load_explicit(&arr[1], std::memory_order_relaxed);
    int v2 = std::atomic_load_explicit(&arr[2], std::memory_order_relaxed);
    std::atomic_thread_fence(std::memory_order_acquire);
//  v0, v1, v2 might turn out to be -1, some or all of them.
//  Otherwise it is safe to read the non-atomic data because of the fences:
    if (v0 != -1)
        print(data[v0]);
    if (v1 != -1)
        print(data[v1]);
    if (v2 != -1)
        print(data[v2]);
}

Example

Scan an array of mailboxes, and process only the ones intended for us, without unnecessary synchronization. This example uses atomic-fence synchronization.

const int num_mailboxes = 32;
std::atomic<int> mailbox_receiver[num_mailboxes];
std::string mailbox_data[num_mailboxes];
// The writer threads update non-atomic shared data 
// and then update mailbox_receiver[i] as follows:
mailbox_data[i] = ...;
std::atomic_store_explicit(&mailbox_receiver[i], receiver_id, std::memory_order_release);
// Reader thread needs to check all mailbox[i], but only needs to sync with one.
for (int i = 0; i < num_mailboxes; ++i)
    if (std::atomic_load_explicit(&mailbox_receiver[i],
        std::memory_order_relaxed) == my_id)
    {
        // synchronize with just one writer
        std::atomic_thread_fence(std::memory_order_acquire);
        // guaranteed to observe everything done in the writer thread
        // before the atomic_store_explicit()
        do_work(mailbox_data[i]);
    }

See also