MobileRead Forums - View Single Post

NiLuJe · 08-23-2013, 11:14 AM

Apples vs. oranges.

You're comparing a snapshot from a point in time where the framework is doing background tasks (catalog/cleanup/scan/indexing) [mostly from some of the niced threads, hence most of the CPU usage falling in the 'nice' column], to one where it isn't.

That's not even taking into account that the CPU frequency is variable, so, depending on which state the CPU governor thought the CPU should be in, top's percentage of CPU usage for the same task can vary wildly.

08-23-2013, 11:14 AM	#2
NiLuJe BLAM! Posts: 13,506 Karma: 26047202 Join Date: Jun 2010 Location: Paris, France Device: Kindle 2i, 3g, 4, 5w, PW, PW2, PW5; Kobo H2O, Forma, Elipsa, Sage, C2E	Apples vs. oranges. You're comparing a snapshot from a point in time where the framework is doing background tasks (catalog/cleanup/scan/indexing) [mostly from some of the niced threads, hence most of the CPU usage falling in the 'nice' column], to one where it isn't. That's not even taking into account that the CPU frequency is variable, so, depending on which state the CPU governor thought the CPU should be in, top's percentage of CPU usage for the same task can vary wildly. Last edited by NiLuJe; 08-23-2013 at 11:17 AM.