For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /max/get-started.md).

Mojo function

sync_parallelize

def sync_parallelize[origins: OriginSet, //, func: def(Int) raises capturing thin -> None](num_work_items: Int, ctx: Optional[DeviceContext] = None)

Executes func(0) ... func(num_work_items-1) as parallel sub-tasks, and returns when all are complete.

TODO: Currently exceptions raised by func will cause a trap rather than be propagated back to the caller.

Parameters:

origins (OriginSet): The capture origins.
func (def(Int) raises capturing thin -> None): The function to invoke.

Args:

num_work_items (Int): Number of parallel tasks.
ctx (Optional[DeviceContext]): Optional CPU DeviceContext to execute the tasks on.

def sync_parallelize[FuncType: def(Int) -> None](func: FuncType, num_work_items: Int, ctx: Optional[DeviceContext] = None)

Executes func(0) ... func(num_work_items-1) as parallel sub-tasks, and returns when all are complete.

Parameters:

FuncType (def(Int) -> None): The body function type.

Args:

func (FuncType): The closure carrying the captured state of the body function.
num_work_items (Int): Number of parallel tasks.
ctx (Optional[DeviceContext]): The CPU DeviceContext to enqueue the work on.