I understand that the Haskell runtime creates an OS thread on each core or so. Workflow threads on these pre-deployed OS threads are then scheduled threads / user threads. Rough.
But how is the Haskell runtime structured - is it monolithic or consists of separate parts?
What I really want to know is how scalable and distributed RT itself is across all cores - for example, can I theoretically have a 1000-core system and run Haskell RT on 5% of the cores (either 50 independent or distributed RT)?
source share