Advanced

Untitled

Story: About the Kernel 3.0 "Power Regression" Myth Total Replies: 3
Author Content
gus3

Sep 30, 2011
6:49 PM EDT 		The entire article left a bad taste in my mouth, for two primary reasons:

1. The power consumption of a totally idle system borders on useless as a measurement. Only monitor systems with alarms are useful when totally idle; desktops, laptops, and netbooks, are simultaneously idle and useful only until the screen's contents are read.

The whole purpose of a personal computer is to react to me when I press a key or move the mouse.(*) If I could type for four hours without charging, before a kernel update, but now after said update I can type for only 90 minutes before the red light comes on, then something has started consuming more power per keystroke, or generating more internal events. Eliminating the UI libraries as a culprit leaves two suspects: the application itself, and/or the kernel.

2. The tone of blaming the ignorant users for reporting their observations, reminds me of an episode not many years back, in which a prominent Desktop Environment's lead developers dismissed out-of-hand the lamentations of the unwashed masses.

(*)I am purposely omitting media playback, which totally alters the base conditions of power use testing.
fewt

Sep 30, 2011
7:17 PM EDT 		The interrupts of an idle system are important, they establish a baseline. You can't create a baseline of a system that is busy, unless you can completely re-create the scenario.

Yes, I completely agree that a computer's job is to react when you provide input, but then the computer isn't idle. If a computer isn't idle, the CPU isn't in a sleep state.

If it isn't in a sleep state, it's awake. The more things keeping it awake, the more wakeups powertop reports.

There isn't any solid evidence that a computer's battery life has decreased after a kernel update, that's the problem. I'm running 3.0.3, and I still get 11 hours, just as I did with 2.6.35.

My Sandy Bridge i5 Thinkpad that arrived today reports 6 hours with WIFI on. Of course, my distribution is tuned for battery life because that is important to me (and users of my distribution).

As hard as it is to swallow, there is no problem, except PEBKAC.
gus3

Sep 30, 2011
7:29 PM EDT 		As an aside, the title of this thread was supposed to be:

"you're doing it wrong" is not a solution
fewt

Sep 30, 2011
7:39 PM EDT
Quoting:"you're doing it wrong" is not a solution


I agree, that's why I also instructed people how to properly test.

Post a reply to this thread:

A free account is needed to post. Write your comment here first: your text is kept in this browser while you sign up or log in, and is waiting for you when you come back.

Your message:
|||||
10
Markdown: **bold**, *italic*, [text](https://link), - lists, > quotes, $x^2$ for math. Formatting help

Sign up and post (it is free)  or  Log in and post