lockfree

lockfree is a collection of lock-free data structures written in standard C++11 and suitable for all platforms - from deeply embedded to HPC.

What are lock-free data structures?

Lock-free data structures are data structures that are thread and interrupt safe for concurrent use without having to use mutual exclusion mechanisms. They are most useful for inter process communication and often scale much better than lock-based structures with the number of operations and threads.

Why use lockfree

• Written in standard C++11, compatible with all platforms supporting it
• All data structures are thread and interrupt safe in their respective use cases
• No dynamic allocation
• Optimized for high performance
• MIT Licensed
• Additional APIs for newer C++ versions

What data structures are available?

Single-producer single-consumer data structures:

• Queue - Best for single element operations, extremely fast, simple API consisting of only 2 methods.
• Ring Buffer - A more general data structure with the ability to handle multiple elements at a time, uses standard library copies making it very fast for bulk operations.
• Bipartite Buffer - A variation of the ring buffer with the ability to always provide linear space in the buffer, enables in-buffer processing.
• Priority Queue - A Variation of the queue with the ability to provide different priorities for elements, very useful for things like signals, events, and communication packets.

Note: These data structures are more performant and should generally be used whenever there is only one thread/interrupt pushing data and another one retrieving it.

Multi-producer multi-consumer data structures:

• Queue - Best for single element operations, extremely fast, simple API consisting of only 2 methods.
• Priority Queue - A Variation of the queue with the ability to provide different priorities for elements, very useful for things like signals, events, and communication packets.

Note: These data structures are more general, supporting multiple producers and consumers at the same time. However, they have storage and performance overhead compared to single producer single consumer data structures. They also require atomic instructions which can be missing from some low-end microcontrollers.

How to get

There are three main ways to get the library:

1. Using CMake FetchContent()
2. As a git submodule
3. By downloading a release from GitHub

Configuration

lockfree uses cacheline alignment for indexes to avoid the False Sharing phenomenon by default, avoiding the performance loss of cacheline invalidation on cache coherent systems. This aligns the indexes to...

FAQ

Why would I use this over locking data structures on a hosted machine?

The biggest reason you would want to use a lock-free data structure in such a scenario would be performance. Locking has a non-negligible runtime cost on hosted systems as every lock requires a syscall. Additional benefits would be performance from cache locality as lock-free data structures are array-based or code portability to non-POSIX environments.

Why use this over RTOS-provided IPC mechanisms on an embedded system?

While locking usually isn't expensive on embedded systems such as microcontrollers, there is a wide variety of RTOS-es and no standardized API for locking. lockfree provides a way to build portable embedded code with a negligible performance cost as opposed to locking. Code using lockfree can be compiled to run on any embedded platform supporting C++11. Additionally, the code can easily be tested on a host machine without the need for mocking.

What advantages does using C++ over C provide for the library?

• Type safety, as data structures are type and size templated
• Much simpler and less error-prone instantiation
• Higher performance due to compile-time known size and header-only implementation
• Encapsulation, the data buffer is a class member instead of being passed by a pointer

What is the formal classification of the data