Re: Processor affinity
Did quite some affinity experiments using latency monitor. Measurements showed that in cores 0-1, IRQ/DPC max-times were the longest.
Concluded that CPU spiking improved a little by forcing PT to execute on higher numbered cores 2-12, leaving cores 0-1 to whatever windowsy, goes on there.( i7 8750H OMEN laptop 6 real cores 12 virtual cores)
A greater difference came out of inactivating "Control method battery" in device manager. That broke the battery so don't do that. Article found at Sweetwater.
But now for some reason I don't know, perhaps some update, it seems to have improved and I don't usually do affinity. Yet, I would try it again if problems occurr.
Skickat från min Moto G (5) via Tapatalk
__________________
Free beer tomorrow !!
- but until then -
we're stuck in the present moment
|