Keeping your memory in order

December 4, 2008

Here is a short description of the different memory ordering for concurrency in C++0x; Bartosz Milewski’s blogs port about C++ atomics and memory ordering.

The default for atomic variables is the familiar sequential ordering (memory_order_seq_cst), its also the slowest as it inserts lots of most memory fences that prohibit the CPU from reordering code as it sees fit. Between that and no ordering (memory_order_relaxed), there are also ordering that just insert fences on reads (memory_order_acquire), on writes (memory_order_release), and on both read and write (memory_order_acq_rel). Being able to choose a memory order other than the safe and slower default is a nice feature, and corresponds well with the C++ deign rationale of not having to pay for what you don’t use. The java volatile’s apparently works much like the c++0x atomic’s with the default sequential ordering, except java volatile’s are not atomic (there is a java library for that).


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: