What's the benefit of using scope over in_place_scope?

[vdrn]

As per docs of in_place_scope:

This is just like scope() except the closure runs on the same thread that calls in_place_scope(). Only work that it spawns runs in the thread pool.

So my question is why wouldn't you want this behavior always? What's the benefit of scope not running the closure in the thread that it was called from?

[cuviper]

• If your pool is fully using all OS concurrency, then running the closure in an outside thread may be over-subscribing, especially if it's doing a lot of work on its own.
• The closure may benefit from the pool's configured stack size, or associating itself with the pool's custom thread names for external debuggers, etc.
• It may be faster to queue work from within the pool, although this is hard to quantify.
• For ThreadPool::scope, it ensures that all contextual rayon calls (like parallel iterators) use that pool too.