Memory Synchronization Intrinsics
Concurrent access to a variable shared by multiple threads always needs to be synchronized, not just in multi-host or multi-node systems.
Intel architecture processors support synchronization by atomic operations on memory. These operations are wrapped into functions supported by the compiler and work across different operating systems. The function generates a full memory barrier (or fence) to ensure that memory operations are completed in order.
Parameters passed to exchange, increment and decrement functions must be aligned on a 64-bit boundary.
With INtime for Windows, you can use Interlock* functions to synchronize access to one variable used in INtime and Windows.