It’s definitely something you need to take into consideration in the design. There are really two options here – and the best depends a bit on how you’d want your library to be used.

If the library provides a high level of abstraction, you might want to use the TPL within your library to allow the performance gains available to surface. However, if its a lower level API, it could easily be better to design your library to be thread safe, so the caller can parallelize as needed.

-Reed

Partition your problem into enough tasks to keep each processor busy throughout the operation, but not more than necessary to keep each processor busy.
Partition your problem in a way to place the most work possible into each task.

So if I am making a library for re-use, are their any design considerations I need to take into account before using parallelism? For example, if I am using my library to accomplish some tasks and I feel that doing these tasks in parallel would improve performance, will I experience a performance decrease if I use TPL because of my library?

Thanks,
Chris