nlcpy.request.flush

nlcpy.request.flush(VENode venode=None, bool sync=True)

Flushes stacked requests on VH to specified VE.

Parameters:

venodeVENode

Targe VE node. If set to ‘None’, the current active VE node is set.

syncbool

Whether synchronize completion of VE execution. Defaults to True.

Examples

>>> import nlcpy as vp
>>> ve0 = vp.venode.VE(0)
>>> # First, none request on VH and VE
>>> ve0.status['stacked_request_on_VH'], ve0.status['running_request_on_VE']
(0, 0)
>>> # Execute some function
>>> _ = vp.arange(10)
>>> # A request is stacked on VH
>>> ve0.status['stacked_request_on_VH'], ve0.status['running_request_on_VE']
(1, 0)
>>> # Flush requests into VE#0 without synchronize
>>> vp.request.flush(venode=vp.venode.VE(0), sync=False)
>>> # Requests are running on VE
>>> # (write_mem request for function arguments and arange kernel request)
>>> ve0.status['stacked_request_on_VH'], ve0.status['running_request_on_VE']
(0, 2)
>>> # Flush requests into VE#0 with synchronize
>>> vp.request.flush(venode=vp.venode.VE(0), sync=True)
>>> # None request on VH and VE
>>> ve0.status['stacked_request_on_VH'], ve0.status['running_request_on_VE']
(0, 0)