We don't intend to provide support for parallelism within the library itself. A parallel library would require a completely different design and would carry overhead that other applications do not need.