On 1 Sep 2014 at 15:27, Dean Michael Berris wrote:
Do you intend this to be a separate library, or would it go into something like Boost.Container?
Not sure yet. My next goal is no-alloc promise-future, also based on spinlocks. I don't intend to "finish" concurrent_unordered_map past what is already done, and I suspect no Boost library will accept it until I or someone else does. I do intend to develop a partial C++ 11 STL to Boost shim library, so one no longer needs the rest of Boost to use Boost.Thread standalone for example. That way we can push Boost.Thread ahead without needing Boost.
You and others may find a description of the algorithm at https://plus.google.com/109885711759115445224/posts/KhbHhU5J19B of use.
Awesome, I'll give that a read. I suspect this is very nice to dig into because it's using the new hotness (transactional memory, and C++11 atomics).
I originally started expecting to make great use of Intel TSX. Read this list's archives for how badly that fared. Niall -- ned Productions Limited Consulting http://www.nedproductions.biz/ http://ie.linkedin.com/in/nialldouglas/