All - For a while we've supported the TBB 'task-based' parallelism, whereby a loop is executed across multiple threads and the results are combined in some way.
As of yesterday I've added support for the other multithreading use case, asynchronous execution of an operation that would otherwise block. This is commonly used in GUI programming but not needed as frequently in scientific computing. Still, there are some use cases where I've wanted it, and now it is available. The first such occurrence is overlapping disk I/O with MPI communication. For some background on the important ideas see here: http://www.devx.com/SpecialReports/Article/38883 for the actual implementation, I've introduced Threads::Thread. This is a typedef to std::thread where available, falling back to a typedef to tbb::tbb_thread where available, or a simple shim class otherwise. The only thing to watch out for is that the tbb::tbb_thread implementation is incomplete in that it does not support assignment, so while strictly std::threat t = std::thread(foo); will work, tbb::tbb_thread t = tbb::tbb_thread(foo); will not. Therefore I'm generally inclined to use AutoPtr<Threads::Thread> to overcome this limitation where assignment is desireable. Note that boost::thread is also a workable alternative here, and I'll probably add support for it too along the lines of 'BEST_STD_THREAD'… Of course, this programming model introduces a whole new opportunity for Heisenbugs, so please report any issues and be careful out there! -Ben ------------------------------------------------------------------------------ Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_d2d_nov _______________________________________________ Libmesh-devel mailing list Libmesh-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/libmesh-devel
