The Concurrent Data Structures (CDS) library is a collection of concurrent containers that don't require external (manual) synchronization for shared access, and safe memory reclamation (SMR) algorithms like Hazard Pointer and user-space RCU that is used as an epoch-based SMR.
CDS is mostly header-only template library. Only SMR core implementation is segregated to .so/.dll file.
The library contains the implementations of the following containers:
- lock-free stack with optional elimination support
- several algo for lock-free queue, including classic Michael & Scott algorithm and its derivatives, the flat combining queue, the segmented queue.
- several implementation of unordered set/map - lock-free and fine-grained lock-based
- flat-combining technique
- lock-free skip-list
- lock-free FeldmanHashMap/Set Multi-Level Array Hash with thread-safe bidirectional iterator support
- Bronson's et al algorithm for fine-grained lock-based AVL tree
Generally, each container has an intrusive and non-intrusive (STL-like) version belonging to cds::intrusive and cds::container namespace respectively.
Version 2.x of the library is written on C++11 and can be compiled by GCC 4.8+, clang 3.6+, Intel C++ 15+, and MS VC++ 12 (2013) Update 4 and above
Download the latest release from http://sourceforge.net/projects/libcds/files/
See online doxygen-generated doc here: http://libcds.sourceforge.net/doc/cds-api/index.html
Evolution of libcds (Gource visualization by Landon Wilkins): https://www.youtube.com/watch?v=FHaJvVdmJ0w
How to build
- *nix: use CMake
- Windows: use MS Visual C++ 2015 project
Some parts of libcds may depend on DCAS (double-width compare-and-swap) atomic primitive if
the target architecture supports it. For x86, cmake build script enables -mcx16 compiler flag that
switches DCAS support on. You may manually disable DCAS support with the following command line flags
in GCC/clang (for MS VC++ compiler DCAS is not supported):
-DCDS_DISABLE_128BIT_ATOMIC- for 64bit build-DCDS_DISABLE_64BIT_ATOMIC- for 32bit build
All your projects AND libcds MUST be compiled with the same flags - either with DCAS support or without it.
Pull request requirements
- Pull-request to master branch will be unconditionally rejected
- integration branch is intended for pull-request. Usually, integration branch is the same as master
- dev branch is intended for main developing. Usually, it contains unstable code
Stack
- TreiberStack: [1986] R. K. Treiber. Systems programming: Coping with parallelism. Technical Report RJ 5118, IBM Almaden Research Center, April 1986.
- Elimination back-off implementation is based on idea from [2004] Danny Hendler, Nir Shavit, Lena Yerushalmi "A Scalable Lock-free Stack Algorithm" pdf
- FCStack - flat-combining wrapper for std::stack
Queue
- BasketQueue: [2007] Moshe Hoffman, Ori Shalev, Nir Shavit "The Baskets Queue" pdf
- MSQueue:
- [1998] Maged Michael, Michael Scott "Simple, fast, and practical non-blocking and blocking concurrent queue algorithms" pdf
- [2002] Maged M.Michael "Safe memory reclamation for dynamic lock-free objects using atomic reads and writes" pdf
- [2003] Maged M.Michael "Hazard Pointers: Safe memory reclamation for lock-free objects" pdf
- RWQueue: [1998] Maged Michael, Michael Scott "Simple, fast, and practical non-blocking and blocking concurrent queue algorithms" pdf
- MoirQueue: [2000] Simon Doherty, Lindsay Groves, Victor Luchangco, Mark Moir "Formal Verification of a practical lock-free queue algorithm" pdf
- OptimisticQueue: [2008] Edya Ladan-Mozes, Nir Shavit "An Optimistic Approach to Lock-Free FIFO Queues" pdf
- SegmentedQueue: [2010] Afek, Korland, Yanovsky "Quasi-Linearizability: relaxed consistency for improved concurrency" pdf
- FCQueue - flat-combining wrapper for std::queue
- VyukovMPMCCycleQueue Dmitry Vyukov (see http://www.1024cores.net)
Deque
- flat-combining deque based on stl::deque
Map, set
- MichaelHashMap: [2002] Maged Michael "High performance dynamic lock-free hash tables and list-based sets" pdf
- SplitOrderedList: [2003] Ori Shalev, Nir Shavit "Split-Ordered Lists - Lock-free Resizable Hash Tables" pdf
- StripedMap, StripedSet: [2008] Maurice Herlihy, Nir Shavit "The Art of Multiprocessor Programming"
- CuckooMap, CuckooSet: [2008] Maurice Herlihy, Nir Shavit "The Art of Multiprocessor Programming"
- SkipListMap, SkipListSet: [2008] Maurice Herlihy, Nir Shavit "The Art of Multiprocessor Programming"
- FeldmanHashMap, FeldmanHashSet: [2013] Steven Feldman, Pierre LaBorde, Damian Dechev "Concurrent Multi-level Arrays: Wait-free Extensible Hash Maps". Supports thread-safe bidirectional iterators pdf
Ordered single-linked list
- LazyList: [2005] Steve Heller, Maurice Herlihy, Victor Luchangco, Mark Moir, William N. Scherer III, and Nir Shavit "A Lazy Concurrent List-Based Set Algorithm" pdf
- MichaelList: [2002] Maged Michael "High performance dynamic lock-free hash tables and list-based sets" pdf
Priority queue
- MSPriorityQueue: [1996] G.Hunt, M.Michael, S. Parthasarathy, M.Scott "An efficient algorithm for concurrent priority queue heaps" pdf
Tree
- EllenBinTree: [2010] F.Ellen, P.Fatourou, E.Ruppert, F.van Breugel "Non-blocking Binary Search Tree" pdf
- BronsonAVLTreeMap - lock-based fine-grained AVL-tree implementation: [2010] Nathan Bronson, Jared Casper, Hassan Chafi, Kunle Olukotun "A Practical Concurrent Binary Search Tree" pdf
SMR
- Hazard Pointers
- User-space RCU
Flat Combining technique
- [2010] Hendler, Incze, Shavit and Tzafrir "Flat Combining and the Synchronization-Parallelism Tradeoff" pdf