Hi.
I have a question concerning threads in a contentprocessor.
Is it thread safe to use some sort of map reduce to accelerate the computations on huge models ?
At least is it possible to use Parallel.Invoke on a mesh to traverse meshparts a litle faster : with xna i had a lot of huge models and lots of submeshes to process and it took about 15mins to build without compression to speed it up. On an i7 @ 2.8ghz up to 3+ghz.
You would have to take the usual precautions when multithreading, such as synchronising access to the map that is common to all threads. There shouldn’t be anything else stopping you from multithreading your content processor.
As I work with threads syncing all day long, it should be easy and work if, as you said, i keep an eye on the synchronisation of the stuff.
Thanks