Does an optimistic lock-free FIFO queue implementation exist?

Question

Is there any C++ implementation (source codes) of \"optmistic approach to lock-free FIFO queues\" algorithm?

Answer 1

Here is my implementation of a lock-free FIFO.

Make sure each item of T is a multiple of 64 bytes (the cache line size in the Intel CPUs) to avoid false sharing.

This code compiles with gcc/mingw and should compile with clang. It's optimized for 64-bit, so to get it to run on 32-bit would need some refactoring.

https://github.com/vovoid/vsxu/blob/master/engine/include/vsx_fifo.h

vsx_fifo my_fifo;

Sender:

my_struct my_struct_inst;
... fill it out ...
while (!my_fifo.produce(my_struct_inst)) {}

Receiver:

my_struct my_struct_recv;
while(my_fifo.consume(my_struct_recv)) 
{ 
  ...do stuff...
}