All about Async/Await, System.Threading.Tasks, System.Collections.Concurrent, System.Linq, and more…
(This answer is based on the .NET Framework 4. As the details below are undocumented implementation details, they may change in future releases.)
No. All of the collections in the new System.Collections.Concurrent namespace employ lock-free techniques to some extent in order to achieve general performance benefits, but traditional locks are used in some cases.
It’s worth noting that purely relying on lock-free techniques is sometimes not the most efficient solution. When we say “lock-free,” we mean that locks (in .NET, traditional mutual exclusion locks are available via the System.Threading.Monitor class, typically via the C# “lock” keyword or the Visual Basic “SyncLock” keyword) have been avoided by using memory barriers and compare-and-swap CPU instructions (in .NET, “CAS” operations are available via the System.Threading.Interlocked class).
ConcurrentQueue<T> and ConcurrentStack<T> are completely lock-free in this way. They will never take a lock, but they may end up spinning and retrying an operation when faced with contention (when the CAS operations fail).
ConcurrentBag<T> employs a multitude of mechanisms to minimize the need for synchronization. For example, it maintains a local queue for each thread that accesses it, and under some conditions, a thread is able to access its local queue in a lock-free manner with little or no contention. Therefore, while ConcurrentBag<T> sometimes requires locking, it is a very efficient collection for certain concurrent scenarios (e.g. many threads both producing and consuming at the same rate).
ConcurrentDictionary<TKey,TValue> uses fine-grained locking when adding to or updating data in the dictionary, but it is entirely lock-free for read operations. In this way, it’s optimized for scenarios where reading from the dictionary is the most frequent operation.