Can `Dagger.@spawn` be used instead of `@async`?

[schlichtanders]

The current documentation of @aync includes this warning

Warning

It is strongly encouraged to favor Threads.@spawn over @async always even when no parallelism is required especially in publicly distributed libraries. This is because a use of @async disables the migration of the parent task across worker threads in the current implementation of Julia. Thus, seemingly innocent use of @async in a library function can have a large impact on the performance of very different parts of user applications.

Hence a natural question is whether Dagger.@spawn can also be used as a better alternative of @async. This would be so cool, as the whole concurrence interface would be unified.

If this indeed works, it would be great to have this somewhere obvious in the documentation.

[jpsamaroo]

In certain cases, yes, you can use Dagger.@spawn like either Threads.@spawn and @async, as it can express a function executed asynchronously either on the same thread, or a different one. However, the exact semantics differ in significant ways. Some of those differences are:

• Dagger.@spawn synchronization is performed implicitly, based on the passed arguments and any surrounding "task queues" that modify task synchronization behavior
• Dagger tasks are generally expected to not mutate or otherwise manipulate their arguments (unless using an API like Dagger.spawn_datadeps
• Dagger tasks are expected to always terminate, and generally should not loop forever, as that may block progress of other tasks (even if the Dagger task yields)
• Dagger will generally only run one task per thread, whereas Threads.@spawn and @async run multiple tasks per thread
• etc.