http://lwn.net/Articles/565715/ This is a special feed containing comments posted to the individual LWN article titled "The 3.12 merge window opens". hourly 2 http://lwn.net/Articles/566127/rss 2013-09-10T02:25:25+00:00 jzbiciak <P>I guess this is a hair-split: If I recall the original article correctly, this work reduces the performance impact of the full-system-idle detection step, so, in <I>that</I> sense improves performance. As a result, because it's cheap enough, it becomes reasonable to detect when the whole system is idle so that that last processor can shut its tick off.</P> <P>If I understood correctly, without this optimization, you would either need to keep the last processor on, or pay a performance penalty.</P> <P>Or am I oversimplifying / misunderstanding?</P> http://lwn.net/Articles/566025/rss 2013-09-07T08:14:08+00:00 chantecode <div class="FormattedComment"> May I suggest a correction:<br> <p> "This work enables the kernel to detect when the entire system is idle and turn off the clock tick, thus improving the performance when the full dynamic tick feature is used. "<br> <p> The goal of this feature is actually not to improve the performances but to improve powersaving. The cpu that handles timekeeping wasn't allowed to shutdown its tick until now. With this patch now it can.<br> <p> -Frederic.<br> </div>