What is a memory fence?

by

For performance gains modern CPUs often execute instructions out of order to make maximum use of the available silicon (including memory read/writes). Because the hardware enforces instructions integrity you never notice this in a single thread of execution. However for multiple threads or environments with volatile memory (memory mapped I/O for example) this can lead to unpredictable behavior.

A memory fence/barrier is a class of instructions that mean memory read/writes occur in the order you expect. For example a ‘full fence’ means all read/writes before the fence are comitted before those after the fence.

Note memory fences are a hardware concept. In higher level languages we are used to dealing with mutexes and semaphores – these may well be implemented using memory fences at the low level and explicit use of memory barriers are not necessary. Use of memory barriers requires a careful study of the hardware architecture and more commonly found in device drivers than application code.

The CPU reordering is different from compiler optimisations – although the artefacts can be similar. You need to take separate measures to stop the compiler reordering your instructions if that may cause undesirable behaviour (e.g. use of the volatile keyword in C).

More Related Contents:

  1. SSE instructions: which CPUs can do atomic 16B memory operations?
  2. Is x86 CMPXCHG atomic, if so why does it need LOCK?
  3. What is the difference between lock, mutex and semaphore?
  4. What is the Swift equivalent to Objective-C’s “@synchronized”?
  5. Would you explain lock ordering?
  6. What is a coroutine?
  7. How does LMAX’s disruptor pattern work?
  8. How do goroutines work? (or: goroutines and OS threads relation)
  9. anonymous struct and empty struct
  10. How to configure concurrency in .NET Core Web API?
  11. Are POSIX’ read() and write() system calls atomic?
  12. What exactly does runtime.Gosched do?
  13. ExecutorService that interrupts tasks after a timeout
  14. Is there a way for multiple processes to share a listening socket?
  15. Django: How can I protect against concurrent modification of database entries
  16. Confused about UPDLOCK, HOLDLOCK
  17. Solving embarassingly parallel problems using Python multiprocessing
  18. Why is (or isn’t?) SFENCE + LFENCE equivalent to MFENCE?
  19. What are the advantages of using an ExecutorService?
  20. NSPersistentContainer concurrency for saving to core data
  21. Memory effects of synchronization in Java
  22. How to make BackgroundWorker return an object
  23. Writing long and double is not atomic in Java?
  24. Do I need to be concerned with race conditions with asynchronous Javascript?
  25. How do I atomically increment a variable in Swift?
  26. Why are two AtomicIntegers never equal?
  27. Java, divide incoming work uniformly via hashing in multithreaded evnironments
  28. c++11 std::async doesn’t work in mingw
  29. Futures / Success race
  30. Singleton double-check concurrency issue

Leave a Comment